CLEANING SYMBIOSIS AMONG CALIFORNIA INSHORE FISHES '
EDMUND S. HOBSON'
ABSTRACT Cleaning symbiosis among shor e fi shes was studied during 1968 and 1969 in southern California, with work centered at La J olla. Three species are habitua l cleaners: the seiiorita, Oxyjulis cali/ornica; the sharpnose seaperch, Phanerodon atripes; and the kelp perch, Brachyistius /1·enatus. Because of specific differences in habitat, there is little overlap in the cleaning areas of these three spe cies. Except for juvenile sha rpnose sea per ch, cleaning is of secondary significance to these species, even though it may be of major significance to certain individuals. The tendency to clean varies between in dividuals. Principal prey of most members of these species are free-living organisms picked from a substrate and from midwater-a mode of feeding that favors adaptations suited to cleaning. Because it is exceedingly abundant in a variety of habitats, the eiiorita is the predominant inshore cleaning fish in Ca lifornia. Certain aspects of its cl eaning relate to the fact that only a few of the many senoritas present at a given time will clean, and that this activity is not centered around well-defined cleaning stations, as has been repor ted for certain cleaning fishes elsewhere. Probably because cleaners are difficult to recognize among the many senoritas that do not clean, other fishes generally do not at tempt to initiate cleaning; r a ther , the activity is consistently initiated by the cleaner itself. An infest ed fish approached by a cleaner genera lly drifts into an unusual attitude that advertises the temporary existence of the transient cleaning station to other fish in need of service, and these converge on the cleaner. Although senoritas, as a group, clean a number of different fishes, a given individual tends to initiate cleaning with member s of j ust one species. The fi shes cleaned most of ten a re those which are most abw1dant and, at the same time, are most heavily infested with external parasites. The most numerous ectoparasites are caligid copepods, the most abundant and widespread of which is Caligus hobsoni. These particular parasites, along with gnathiid isopod larvae, a re the major prey of the cleaning fishes. Cleaning is essentially limited to the external body surface ; ectoparasites of the oral and branchial cavities are not ordinarily taken. Clean ing effectively r educes the number of parasites on fishes that are cleaned, and is an important activity for the organisms involved. However, there is no basis for the co ntention that many good fishing grounds in southern California exist because fi shes have congregated in these locations for cleaning.
It has been suggested that many of the better reasons that have nothing to do with cleaning. inshore fishing spots are, in fact, cleaning sta Regardless of which view is correct in a given tions (Limbaugh, 1961; Feder, 1966) . The situation, one having witnessed fishes crowded contention is that fishes congregate at these lo around a cleaner, vigorously soliciting its ser cations so that ectoparasites and other deleteri vices, can only conclude that this activity is in ous material can be removed from their bodies deed important to the organisms involved. by resident cleaning organisms. Critics of this Cleaning symbiosis has been widely described hypothesis might well suggest instead that clean in the literature (Longley and Hildebrand, 1941; ers simply are especially active where fi shes are Eibl-Eibesfeldt, 1955 ; Limbaugh, 1955, 1961; most abundant, or that the cleaners as well as Randall, 1958, 1962; and others) and was re viewed by Feder (1966). Youngbluth (196 ) those they clean occur at these locations for studied activity of the Hawaiian cleaning labrid Labl'oides phthil'ophagus in some detail, and 1 Contribution of Scripps Institution of Oceanography. • National Marine Fisheries Service, Tiburon Sport Losey (1971) analyzed the communicative sig Fisheries Marine Laborator y, Tiburon, Calif., and nals between this same pecies and the fishes that Scripps Institution of Oceanogr aph y, University of Cal ifornia, San Diego, Calif. Mailing address: National it cleans. But most other reports on cleaning Marine Fisheries Service, Fisher y-Oceanography Center, have been simple treatments based largely on P.O. Box 271, La J olla, Calif. 92037.
Manuscript accepted March 197 1. FISHERY BULLETIN : VOL. 69. NO.3. 1971. 491 FI IIFRY OLLLETJ. ' VOL 69, 0 3 incidental obse]'Yatinns, In this report, I d - scribe cleaning symbiosis among- inshore flshes of southern California ,md attempt to relate oh Mosl of th cleaning during this "en'ed actiyity with the incidence of speciflc ::;luciy \\ a, p I'form d by th fi rita Figure 1) , ectoparasi tes, \\ hich \)y virtu of it..'i great abundance in a va Conrad Limbaug-h, Scripps Institution of I'i ty of habitat." i th PI' d mine nl cleaner in Oceal1ogra phy. was a mon g- the fI rs t to report ::;1101', The 'harpno." seap rch (Figur 2 ) was cleaning s~'mbiosis among California fi shes, In frequently ok I'V ct cl aning, but it acti ity is a stud~' of fishes of the kelp hed::;. Limhaug-h centered ill de p r water. Th k II> perch (Fig (l!););) de::;c rihed cleaning' h~' the seiiorita, (1.1' ure :1) may h an important clean r in the can !ljlliis ('oli/url/inf, a fish of the famlly Lahridae, opy reg-ion of th k Ip for st . wh re it concen- and also h~' seyeral seapprches of the family
Emhiotocidac: the kelp perch, n/'(/( hyistiliR frcl/utlls; the hlack perch. EIII/,int()('11 j((('ks()lIi; and the pile perch. Rh(l('rl('h illls I'ff(,C([, ,uh::; quent ohsen'ers ha\ e de::;Cl'iled cll>an1l1~ hy the rainlJo\\' seaperch, H!Il)SlIrIlS ('(()lli (;otshall. 1967); the sharpllose seaperch. PIUflll'rnd()1I tIt ril'( S (Cia I'k(>. Flechsi!-!', and (; ri~~. 19(,7; Cotshall. 1%7; Hohson, 1969a); and the black smith. C'hl'olllis IJl/llctil,illllis (Turner. Ehert, and Gi\'en. 1969),
FIGlRE 1.- iiorila,
F IGl' KE 2,-Sha r pnose sea per ch a mong branches of a gor gon ian,
492 HOBSON : CLEA INC SYMBIOSIS trates, but this was not determined in this tudy all pecimen and "'ill b report d in dr>t:lil 1: because observations in the kelp-canopy habitat where in collaboration with R. F. r :.:ey, l O.•. were infrequent. Nevertheless, observations ational l\Iu,eum. The e coli ction: al'o pro were sufficient to recognize the kelp perch as a vided the material for de:scription: of 11 :1' ci .. habitual cleaner. The only other fi sh seen clean of cope pod formerly new to scienc ('r ... y, ing was the white seaperch, Phanerodon fur 1969a, 1969b, 1970; J. Ro, alifornia. t:.lt (;1- catt~, in which cleaning seemed to be onl y an lege, Long Beach, unpublbhed manuscript). d occasional incidental activity. ditional undescrib d species may occur among a number of copepods from these coil ction: METHODS presently und r study by Z. Kabata, Biolog-ical Station, Xanaimo, British Columbia. During 1968 and 1969, I spent more than 103 In addition to a survey of the ctopara.ite., hr underwater directly observing cl eaning and gut content of known cleaning :pecie:, including related activity in California inshor e waters. material from .):1 sei1oritas, ~~l :harpllo. e . ('<1- Also contributing to the study are many in ci perch, and 3 kelp perch, were anal~·zecl. dental observations of cleaning made during Many ectoparasites leave their host \\'h n it other work with California fi shes between 1961 is in difliculty, and some fishes regurgitate their and 1970. stomach contents under stress. To reduce this Supplementing the observations, 421 speci loss, all specimens "'ere indi\'idllally seal'd ill mens of 39 species were collected with spear. plastic bags immediately upon capture, and \\'hile These r epresent mo t of the species common in still underwater. the study area that exceed a length of 100 mm To acquire detailed data on the cleaning' in (all lengths of fishes in this r eport are standar d teraction, a numher of indi\'idllals of cl :lninJ,! length). The ectoparasites were coll ected from species \\'ere kept under surveillance for periods
FIGt 'RE :3.-Kelp perch next to g-iant hlp,
I 3 FISflERY B LLETlN: VOL 6?, r-;o J up to 15 min, and a verbal account of their ac in comparison to some area nearby to the so uth tivity was recorded on tape. The attempted and elsewhere in al ifornia. ther details of standard of 15 min could not be maintained for lhe principal study area will be introduced as all these observations because sometimes con they become pertinent. tact with the fish being watched was lost as the During all observation periods at the La Jolla fish swam among vegetation or other fishes. In tations a record was kept of water temperatures dividuals followed included known cleaners as from surface to bottom, horizontal visibility, and well as others that had nol shown evidence of urge condi tions. cleaning. In monitoring the activity of known cleaners, a record was kept of the time during OB ERVATI which they showed an apparent cleaning inter est in other fish and also the number of cI aning GE ERAL ECOL GY bouts in which they became involved. A clean ing bout is defined here as any cleaning activity eiio rita involving a discrete group of fishes, whether this The , enorita, which attain a length of about group includes one cleaner attending a single 2:)0 mm, i one of the most abundant fi hes in fi sh, or several cleaners attending a c1u tel' of the in shore waters of southel'l1 alifornia, in 40 to 50 fi h. On various occasions I also re cluding the hannel Islands. It occurs from the corded the number of times that the cleaner shoreline to depth exceeding 40 m and is re actually "picked" at the body of another fish, and corded from central alifornia outh to central precisely at what point on the body this action Baja alifol'l1ia, Mexico (Roedel, 1953). An in was directed. habitant of water over rocky substrates and mnong ea weed, the senorita ometime wims Study Areas singly, but more often in groups of from a few to many hundreds of individuals. Like other Observations of cleaning symbiosi were made labrids, it is strictly a diurnal fi h, taking shelter at many locations throughout southel'l1 Califor under cover at night. nia, including the Channel Islands, and at the Coronado Islands, Mexico. However, most of the data were collected during concentrated work Food habits.-The senorita feeds on a variety at La Jolla, Calif. Here, three study sites were of benthic organisms from the surface of both established, each including an area of about algae and rocks. It al 0 feeds heavily in the mid 2 100 m , that lie on a line running northwest off waters, taking small organisms in the plankton, shore from La Jolla Point. loving away from as well as forms that are attached to or encrusted the beach along this line, the first station lies in 3 on drifting algal fragments. All this feeding to 10 m of water about 200 m offshore, the sec is accomplished in a characteristic picking man ond is in 20 to 25 m of water about 700 m off ner, a mode of feeding well suited to its pointed shore, and the third in 30 to 35 m of water about snout and the several long, curved canine teeth 1000 m offshore. The sea floor at all three sta that project forward at the front of each jaw. tions is rocky and il'l'egular, with many crevice To determine the food habits of this fish in and ca\'es. Algae are not conspicuous at the two the study area, 26 specimens, 110 to 195 mm deeper stations, which are similar, but the rocks long, were speared randomly from the population support a heavy growth of gorgonians. On the at large. None of these were cleaning when other hand, the nearshore region of the 3- to 10-m collected. Food items in their stomachs, ranked station is richly carpeted with surfgrass, Phyl as percentage of each item in the entire sample, lospadix, and other parts of the inshore station were as follows: bryozoans encrusted on algae, are forested by large kelps, particularly giant 43 '7c ; caprellid amphipods, 32 % ; fish eggs, 3 co ; kelp, Macl'ocystis, and feather-boa kelp, Egre gammarid amphipod , 2.5 %; unidentified crus gia. However, these large kelps are sparse here tacean fragments, 4 ~c ; and pelecypod mollusks,
494 HOBSON: CLEA ING SYMBIOS IS
2.4 %. A number of items each made up less ditions. These comments on temperature effects than 1 % of the sample, including crab frag are based entirely on casual evaluations of rel ments, gastropod mollusks, pycnogonids, and a ative abundance under varying conditions. gnathiid isopod larva. Unidentified material constituted 16 % of the sample. The gnathiid larva, a single individual from one senorita, was Sharpnose Seaperch the only evidence of ectroparasites among this The sharpnose sea perch is not regarded as a material. common species (e.g., Limbaugh, 1955), but was Limbaugh (1955) stated that senoritas are relatively abundant during this study over rocky omnivorous carnivores which feed on almost any substrates below 20 m in the La Jolla area. It animal material. Quast (1968) concluded that grows to over 200 mm long and is recorded from the principal food s of the senorita are small Bodega Bay, central California (Miller, Gotshall, gastropods and crustaceans associated with al and N itsos, 1965), south to the San Benito gae. Because he found no crabs or pis tal shrimps Islands, Mexico (Roedel, 1953). Most of those in the diet, Quast suggested that bottom feeding observed during this study were juveniles less is infrequent; however, having seen senoritas than about 125 mm long that swam singly or, frequently picking on the bottom, I believe there more often, in small groups of less than 10 indi must be some other re aso~ why these prey are viduals. Adults were seen only occasionally but not taken more often. Size may be a factor, as sometimes swam in larger aggregations. All crabs and pistal shrimps generally are larger and activity observed in these fish occurred during more heavily shell ed than most prey of the daylight. After dark they hover above the sub senorita. strate and are alert, but their activity at this time, if any occurs, was not determined. Movements.-Although individual senoritas may range widely over the bottom in a given Food habits.-This seaperch takes a variety locality, they seem to operate within a restricted of benthic organisms from the urface of rocks, range. Individual fi sh, when followed, always algae, gorgonians, and other benthic substrates. criss-cross back and forth within a defined area. Prey are taken off the bottom in a characteristic Twelve individuals, selected randomly from the picking manner similar to that of the seii.orita. population at large and kept under surveillance However, the seaperch's dentition would seem for 11 to 15 min each, showed no evidence of less specialized for picking than that of the cleaning. senorita; its conic teeth are relatively short and Senoritas are most abundant at the 3- to 10-m straight, and those at the front of the jaws are station and become progressively fewer with in not notably longer than those on the sides, nor creasing depth offshore. Nevertheless, even at do they project forward. the 30- to 35-m station the species was among To investigate the food habits of this fish in the most numerous present. Fluctuations in the study area, 13 individuals, 76 to 170 mm numbers were often apparent with changes in long, were speared randomly from the popula water temperature. Some of the movement is tion at large. None of these were cleaning when vertical. When a layer of colder water msponges, 1 r ( . Unidentified material made up ments in certain members of the population. The 14 ~ ;- of the sample. There was no evidence of numbers present fall off noticeably when temper ectroparasites in this material. One individual atures drop much below 13° C, but at least some had fed heavily and exclusively on plankton senoritas were present no matter what the con- ic copepods, showing that this fish is not
495 FISHERY BULLETIN: VOL. 69, NO. 3 limited to benthic prey. I have found no refer with kelp. Attaining a maximum length of about ences to food habits of this fis h in the litera 150 mm, the kelp perch is recorded from Van ture. couver Island, Canada, south to central Baja California, Mexico (Roedel, 1953 ). The kelp lI1OL'ements.-On the basis of limited obser perch occurs near the rocky bottom at the base vations, these fis h do not seem to move ar omid in of giant kelp, as well as adjacent to the rising their habitat as much as seii.o ritas do. Never kelp stipes, but is most abundant just under the theless, they do show marked inshor e/ offshore kelp canopy, near the water's surface. Typi mo\'ements that may r elate to changing water call y, this fi sh occurs in aggregations of a dozen temperatures. Unli ke the ubiquitous senorita, or more, but larger individuals frequently are this fish occurred in limited numbers that al soli tary, especiall y those near the rocky sea fl oor. lowed assessing relative abundance through ac Most of my observations of kelp perch were made tual counts. It \\'as ne\'er ~ een at the 3- to 10-m outside the La Jolla study area, the majority station but was r easonably abundant (10-20 in around the Channel Islands. dividuals were counted during 15-min periods) on all visits to the 30- to 35-m station. At the Food habits.-This perch feeds in a picking :w-to 25-111 station its appearance was irregular ma nner, simila r to that employed by the senorita and closely foll owed temperature flu ctuations. and sharpnose seaperch. It preys on a variety Generally it \\'as rare or absent at the ~O - to 25-m of organisms from the surface of the surrounding station when bottom temperatures rose much above 13° C, and was pr esent (a maximum of kelp and a l 0 feeds extensively on material sus 10 was seen during a 20-min period ) when the pended in the cUlTent. Its pointed snout and small , upturned mouth, together with a number temperature dropped much belo\\' this level. As of relatively long, curved canine teeth that pro most of the individuals seen were juveniles, a ject forward at the front of each jaw, are well seasonal factor independent of temperature was probably operating her e. N e\'er theless, short suited to its mode of feeding. The dentition of term temperature changes over the critical range this fi sh is simila r to that of the senorita, a fact (approximately 12°_14 ° C at the 20- to 25-m sta also noted by Hubbs and Hubbs (1954). I did tion) were consistently accompanied by the not sample kelp perch from the popUlation at presence or absence of this fi sh. I emphasize la rge for food-habit analysis; all those collected that these assessments of abundance are relative were from known cleaning stations. However, Limbaugh (1955) stated that they feed on small to the numbers of the species regularly present. The seii.orita \\'as ahnlYs more abundant than crustaceans, particularly those that occur on the sea perch at all stations and under all con giant kelp. Quast (1968), who also reported a ditions. Thus whereas the sea perch was con predominantly crustacean diet, with a prepon derance of amphipods, noted that some mollusks sidered abundant during a period in which 15 individuals were seen, at no time did I find so few and bryozoans are taken as well. senoritas present at any of the three La Jolla stations. M01·ements.-Limited observations indicate that aggregations of kelp perch in the canopy, Kelp Perch and close to large rocks, remain relatively stable. Several aggregations that were observed over The kelp perch was not abundant in the La 2 to 3 months did not change appreciably in lo .J olla study area, \\'here it was seen onl y at the in cation or in numbers of individuals. Data on shore station. Its distribution is essentiall y lim this point are scanty, however. ited to the kelp beds, which were not well de At night they hover in the same areas in veloped in the study area at the time of this which they are active in daylight, but their ac ,,·ork. ::-\evertheless, it is very numerous in Cal ti vity at thi time, if any occurs, was not deter ifornia inshore waters that are heavily forested mined.
496 HOBSON: CLEANI NG SYMBIOSIS
CLEANING ACTIVITY OF seven instances were senoritas seen cleaning THE SENORITA members of a mixed-species group. The tabu lation of species cleaned (Table 1) does not in Unlike some other cleaners (see Feder, 1966) , clude the mixed-species groups because in the senoritas do not establish well-defined stations mixed groups it was not determined whether at which they receive other fi shes seeking to be representatives of all species present were ac cleaned. Rather, the senoritas, as they move tually cleaned. All seven mixed groups included over the local area, approach and clean fishes halfmoons, Medialuna californiensis, and one or wherever they encounter them. more fi sh of other species. In four of these, Despite their great abundance, only a small halfmoons were mixed with blacksmiths, in one seQ1'l1ent of the senorita population seems pre they were mixed with opaleyes (Gii'ella nigric di;posed to clean at a given time. The cleaning ans), in one with rubberlip perch (Rhacochilus habit is not limited to any particular stage in to xo tes ) , and in one with both rubberlip perch their life history: cleaning senoritas have in and pile perch. All of these were incidental cluded some of the smallest individuals seen observations. The compilation does not include « 40 mm) as well as some of the largest data obtained on other occasions when the ac ( > 225 mm). In cleaning material from the tivity of individual cleaners was recorded for bodies of fi shes, senoritas employ the same pick extended periods. ing technique they use to take small prey from The data clearly indicate that blacksmiths, and a rock or algal substrate. This mode of feeding, to a lesser extent topsmelt (A therin01)s affinis) , along with their pointed snout and long, for predominate as recipients of the senorita's clean ward-projecting canine teeth, are well suited to ;ng efforts in the areas where the observations the cleaning habit. ,,-e re made. Table 1 is not a definitive list of Individuals that clean are numerous where species cleaned by the sei'iorita; nevertheless, it there are many resident fishes, especially of is evident that man ~- species which co-occur with certain species (as discussed below), but I found the senorita are not cleaned. At other times, no evidence that residents of other areas come in addition to all species noted in Table 1, I have to these locations to have parasites removed. seen Seriola dorsalis and T}"((ch1ll"/{ ~ symmet Occasionally a migrating species, such as the ric1ls being cleaned. But the ratio of species California yellowtail, SeriolcL dorsalis, will pause listed here generally is consistent with observa to be cleaned while passing through areas where tions made on other occasions and at many dif cleaners are active, but this is not the same as ferent locations. a resident of a particular area habitua ll y swim ming elsewhere to be cl eaned and then returning to its home ground. TABLE l.-Fifhes observed being cleaned by senoritas during 62 observation periods between June 1968 and Fishes Cleaned by the Senorita January 1969 at La Jolla, Calif. (exclusive of seven mixed-species groups).
Percent of Casual observation alone show that some fi sh Species toto I bouts species are cleaned f ar more often than others, observed and that m any species do not seem to interact Blacksmith, Ch romis plln ctipinniI 231 60 Topsmelt, .1thtrinopl a[TiniJ 81 21 with cleaners at all. Garibaldi, lfyprYPopJ rubi(unda 22 6 To obtain data on this point, a record was kept Halfmoon, J/ldia/una colijorninuis 19 5 Senorita, Ox\'ju/is {olijorllicQ 10 3 of the species seen being cleaned by senoritas Rubberlip perch, Rhnco(h i/uJ to.\otn 8 2 during 62 observation periods (15 min to 2 hI' Opoleye, Glrt/la nigri.canJ 5 Kelpfish, /lrlaoJ/iehllJ rOJtratUJ 3 long) from June 1968 to J anuary 1969. During Black perch, Em-biotoca jackJoni this period, 392 cleaning bouts were witnessed, Pile perch, RharoehiiuJ 'p oeca Sorgo, .41liJOtrl'mllJ do",idJoni 385 of which involved senoritas cleanin g one or Bl ue rockfish, S,baJtn mYJ/inuJ more individuals of a single species; in only Olive rockfish, StboJltJ urranoidu
497 FISHERY BULLETIN : VOL. 69, NO. 3
Reports in the literature present a comparable Generally the first sign of an interaction oc picture. Most published accounts of cleaning by curs when a senorita swims up alongside a black senoritas describe the way blacksmiths cluster smith in mid water and closely inspects its body. around this cleaner to solicit its attentions (Lim The blacksmi th may then immediately stop swim baugh, 1955, 1961; Feder, 1966; and others). ming and, holding its fins motionless and erect, Limbaugh (1955) observed the following fish drift into an awkward-appearing posture. Usu being cleaned by senoritas: M yliobatis califor ally the 'blacksmith is head-down, but sometimes n ica, Stereolepis gigas, Paralabrax clathratus, turns on its sides or is tail-down. On some occa Trachurlls symmetricus. Atherinops affil1Ul, sions the blacksmith presents a particular part Anisotremus dal'idsol1i, Hyperprosopon argen of its body to the inspecting senorita. The sen tellm. Rhacochilus l'acca, Chromi.s punctipinnis, orita swims about this fish, usually pausing Hypsypops 1"llbicunda, Girella l1igrical1s. M edi briefly to pick at its body. Immediately follow aluna californiensis, and Mola mola. Turner et ing the first sign of this activity other black al. (1969) observed the following fish being smiths converge on the spot, so that very quickly cleaned by senoritas: Seba.s tes spp., Atherinops 10 or more crowd around the cleaner (Figure 4) . affinUl, Atheril1opsl'.s califO?"11iellsis. Trachurus The senorita soon leaves the original blacksmith symmetriclls, Seriol.a dorsalis. Chromis pUl1ctip and may then move on to one of the others. It innis, and Mola mola. either of these reports may also swim slowly away, whereupon the gives data on the relative frequency with which group of blacksmiths follows along, each attempt these different species were cleaned, but it is ing to position itself in the senorita's path. Al significant that many of the same species con though the senor ita shows progressively less in sistently appear in the r eports of independent terest in the blacksmith , they continue to crowd observers, while at the same time many other in its way. Soon the senorita show no further species that frequent these waters in large num interest in cleaning, and all but a few black bers are not mentioned. No doubt many species smiths leave the group. The remaining few not yet reported are occasionally cleaned by doggedly continue attempting to present them senoritas, but there seems little doubt that a selves to the now-unre ponsive cleaner. Even certain few species, the blacksmith in particular, tually, however, the e last blacksmiths lose con predominate in this activity. tact with the cleaner as it swims off among the kelp or the many other senoritas in the sur Specific Cleaning Interactions rounding water. Once they have lost contact with the cleaner the blacksmiths do not attempt The fishes cleaned by the senorita vary ma rk to solicit cleaning from any of the many other edly in their habits and habitat. as well as in senoritas around them. their relative numbers. These fishes do not seek On only two occasions did I note blacksmiths out cleaning at a "station" established by the soliciting cleaning from a senorita that did not sei'iorita. but rather receive the senorita on their seem to have made an initiating gesture. Once own grounds during the course of their regular the blacksmiths were very small, about 40 mm activity. Cleaning interactions often proceed long, and in the other observation, at a depth differently with one of these species than with of 27 m, little cleaning had been seen and rela another. Some of these variations in cleaning tively few senorita were present. However, activity are characterized below. in both instances the senoritas were known by me on the basis of earlier observations to be Senm·ita-blacksmith inte1'actions.-The black individuals that clean. It is possible that the smith is one of the most abundant fish over rocky fishes soliciting attention recognized these sen substrates in California inshore waters, where oritas as cleaners through some cue not noted it swims in large stationary aggregations in mid by me. Sometimes when a senorita incidentally water. It feeds largely on zooplankton (Quast, passes close to a blacksmith, the blacksmith no 1968) and attains a length of about 250 mm. ticeably pauses in its swimming and looks as
498 HOBSON : CLEA. NING SYMBIOS IS
FIGURE 4.-Sefiorita cleaning the caudal peduncle of one of a group of blacksmiths that hover to solicit service. though it is beginning to assume a soliciting at a number of different locations-always with posture; however, when the senorita swims on blacksmiths. I have no explanation for the fact past, the blacksmith immediately resumes its that a seilorita which becomes unresponsive and original activity. OccasiOll all y members of other leaves one group of blacksmiths that sti ll vigor species were seen responding simil arly to passing ously so licits its service may soon initiate ac senoritas. In most observations of cleaning, my tivity again with another blacksmith. attention was drawn to activity al ready in prog The three individuals whose cleaning activity ress, so that it was not possible to determine was monitored for 15 min joined in a mean of whether cleaner or client had initiated the ac 4 separate bouts (range 2-6). For a mean of 11 tivity. min of this time (range 6.75-13.25 min) they Individual senoritas that cleaned blacksmiths showed an apIJa rent cleaning interest in black during many short-term observations were not smiths. or wel·e accompanied by blacksmiths with seen cleaning any other species. This same which they had earlier initiated a cleaning inter situation held true for three individuals, known action. vYhen not thus engaged with black to have been cleaning blacksmiths, whose ac smiths, they s\\"am in midwater showing no tivity was monitored in detail on tape for 15 apparent interest in the fishes around them but min. When observed for extended periods, occasionally picked at drifting scraps of debris, senoritas were found to become in volved in usually algal fragments. During much of the a succession of separate cleaning bouts. This time that they swam in consort ,,"ith blacksmiths, activity was not restricted to one location but they closely inspected these fish and actually continued at various points over a relatively picked at their bodies a mean of 26 times (range wide area" Periodically they j oined cleaning al 14-33) . Of these picks, 27 (;. were made at the ready underway, or initiated cleaning themselves base of the blacksmith's anal fin, 25 (;. on the
499 FISHERY B LLETI : VO L. 69, NO. J caudal peduncle or caudal fin, 22jr at the base of group of topsmelt more readily than they do the pectoral fin, 10 «(- somewhere on the body with blacksmiths, as they n ed onl y swim down exclusive of a fin-base or head, 8jr on the head, 10 the subs1rate below, wher the topsmelt seem 5('~ at the base of the pelvic fin s, and 3jr at the reluctant to follow. base of the dorsal fin. These shallow areas are frequently swept by Clearly, the bases of the fin s receive most of surge, and 1he load of drifting debris in mid the attention from the senoritas. These data water is frequently heavy. In this area cleaning are consistent with the many more general ob se110ri1as frequ nlly leave 1he group of topsmelt servations made on other occasion. At no time they are attending to inspect an object drifting during this study were senoritas seen to clean in the watcr nearby. Som times they take the within the oral or branchial cavities of black object in10 their mou1hs, sometimes not. Often smiths: all cleaning wa directed at the body when taken it is quickly rejected. surface. Attempts at extended ol)servations on individ ual senori1as 1hat had been cleaning topsmelt Seiiorita-topsmelt i/lteractio/ls.-The top were largely unsuccessful. Too often before the smelt, ,\'hich attains a length of about 200 mm, is observa1ion had progressed far the senoritas abundant in many inshore regions of California eli appcared among the surf grass or other vege- coastal waters, but it di stribution is more spotty 1ation carpeting the sea floor in this area. How than that of the ubiquitous blacksmith. Like ever, two individuals were followed for 10 min the blacksmith, it feeds largely on zooplankton, each, during which time one entered into four, which it takes while swimming in large schools the olher two, separate cleaning bouts. Between at the water's urface. Quast (196 ) noted the cleaning bouts these two swam over a wde area similarity in diet between topsmelt and black alone in midwater, occasionally picking a~ drift smiths, and while acknowledging that their feed ing debri. On several occasions they picked at ing areas may overlap, he pointed out that top benthic algae. Neither individual showed clean smelt normally swim higher in the water column. ing interest in ,pecies other than top melt, which In the La Jolla study area, topsmelt are con was consistent with observations of other sen centrated at the inshore station over extensive oritas that cleaned topsmelt. fields of surfgrass that grow in 3 to 5 m of water. They are never far from the substrate Seiiorita-,qaribaldi illteractions.-The gari in this relatively shallow water, even though baldi, which attains a length of about 250 mm, they swim in large schools at the water's ur is a solitary, highly territorial fi h that lives face. They are more abundant than blacksmiths close to the ubstrate. Especially during the re in this area, and here they predomin ate in the productive season, when the males aggressively senorita's cleaning activity. guard their ne ts among the rocks, these bright The cleaning interaction proceeds in much the orange pomacentrids normally drive away all same way as it does with blacksmiths: the ac other fish that come near. They feed on sessile tivity is initiated when a senorita swims up to benthic invertebrates and are abundant at the an individual topsmelt and begins to inspect it 3- to 10-m station. closely. Immediately other topsmelt converge Garibaldis frequently are cleaned by enoritas. on this pair to place themselves in the se110rita's Most of the garibaldis seen being cleaned were path, thus soliciting its attention. When pre swimming a meter or so above the bottom; I senting themselves motionless before sei'ioritas, did not observe cleaners active around the gari topsmelt frequently hover tail-down, in contrast baldis guarding nests among the rocks. All of to the head-down posture most often assumed the garibaldis seen being cleaned were solitary, by blacksmiths. I sm\' senoritas clean only the which reflects their territorial nature. The sen external body surfaces of topsmelt. In the rel orita swims up to a garibaldi and closely inspects atively shallow water where most of this activity its body, thus initiating the action. Usually the was observed, se110ritas break off contact with a garibaldi hovers motionless in a normal hori-
500 HOBSON : CLEANI NG SYMBIOS IS
zontal attitude, its fin s sometimes erect. The not particul arly abundant in the principal study senorita may pick at a few places on the gari areas. Still, they were frequently seen being bald i's body-most often around the caudal re cleaned by senoritas during thi s stUdy. When gion-but usually its attentions are brief, and many halfmoons were present, cleaning by the soon it swims away. With blacksmiths and top senorita progressed much as described above for smelt, each cleaning bout is prolonged by the blacksmiths. Yet when just one halfmoon was many other individuals that join at the cleaning present, a frequent occurrence, the cleaning site to crowd in the senorita's path. Nothing of bouts were brief like those described above for this sort happens with the solitary garibaldi, the garlbaldi. At least one halfmoon was pres which usuall y makes no attempt to follow the ent in all the mixed-species grou ps that I re senorita when it leaves, so that each cl eaning corded when collecting the data presented in bout is relatively brief. After leaving one gari Table 1. I saw senoritas clean only the external baldi, however, often the senorita qui ckly ap body surface of halfmoons. proaches another. In agreement with their One senorita, seen cleani ng a halfmoon, was cleaning of blacksmiths and topsmelt, senoritas kept under surveill ance for 12 min before con known to have cleaned garibaldis were subse tact was lost. As the observation period began, quently seen cleaning only other members of that the senor ita picked at the halfmoon once and same species. This was true during several then moved away, swimming slowly and alone, short-term observations, and also when one in 2 or 3 m over t he substrate. After an unevent dividual was followed for 15 min, and a record ful 3 mi n, the seiiorita approached a second half of its activity was taped; this particular senorita moon, which promptly hovered in a head-down initiated cleaning activity with 26 different gari Il ttitude. For 15 sec the senorita closely in bald is during the observation period as it swam spected this halfmoon and picked at its body over an irregular course among the rocks in an three times before swimming away. It then area where blacksmiths, topsmelt, and other spe continued on alone for the remaining 8+ min cies also were present. Each cleanin g bo ut lasted that it was under obsen'ation, still swimming a mean of 10 sec (range 7-25 sec) , totaling 4 min slow ly over a wide semicircular course 2 or 3 m 15 sec of the 15-min period. In nine of these above the rocks. During this time it passed bouts, the senorita inspected the garibaldi but many diffe rent fi sh wi thout showing interest, did not pick at its body. In the other 17 bouts, but it did not pass another halfmoon. It did the senorita picked at the garibaldi's body a pick at three di ffe rent pieces of fl oating debris total of 42 times, or a mean of about 2.5 times but rejected a ll three immediately. per bout. All cl eaning of garibaldi that I observed was S e ii 0 I' it a-s e ii 0 l' ita intemct iOlls.-Senori tas directed at the external body surface. themsel ves are cleaned by other members of their own species. Despite the large numbers Seiim'ita-halfmoon interactions.-Halfmoons, of senoritas that usua ll y are present, I saw no which may exceed a length of 250 mm, usually groups converging on cleaning individuals, as swim high in the water col umn, frequently in r egul arly occurs with blacksmiths, topsmelt, and large aggregations, but just as often in small other abundant species. In most of the sei'io groups or as soli tary individuals. They are often rita's in traspecific cleaning interactions, the abundant among rising stands of giant kelp. cl eaner attends just a single individual, which Their omnivorous diet, which includes a variety usuall y hovers motionless in a normal horizontal of benthic algae, along with bryozoans, sponges, attitude, except that its fi ns are erected; some and crustaceans (Limbaugh, 1955; Quast, 1968), times the mouth is open wide and gill covers indicates bottom feeding; however, much of this are distended, but I saw senoritas clean only material is taken in midwater as drifting debris. the externa·1 body surface of these fi h. There Considering their large numbers in many was no indication that senoritas which clean southern California coastal areas, halfmoons are other senoritas also cl ean other specie. I fol-
501 FlSIILRY BLLLETIN, VOL 69, NO 1
lowed one individual for 10 min after having passed by without responding to the kelpfish, and seen it clean another sei'iOJ'ita. After this ini each time the kelpfish relurned to the cover be tial activity, the indi\'idual under surveillance low, where it waited until the next approach, swam o\"er a wide area, showing interest only Finally a pa. sing senorita pall ed briefly and in other senoritas. e\ en though blacksmiths, top pickecl at lhe kelpfish's ide b fore continuing on smelt, and other species were present. Swim its way. After this brief encounter, the kelp ming alone, :2 or 3 m o\'er the rocks, it would fish c1id not ri e from concealment again even assume a posItion alongsine another seii.orita and though s veral more enoritas sub equently folIo\\' it for a short distance. l 1sually these passed o\'erhead. other fish sho\\'ed no interest. hut some stopped During the pre' nt study thi sequence of swimming and erected their fins, whereupon the e\'ents in\'ol\'ing kelpfi.h and senoritas was wit cleaner picked at their JH)(iies-usually once, hut nes:ed :everal time.' at a variety of locations; occasionally se\eral times. nf't\\"('ell cleaning indeed, e\'ery in tance of a kelpfish being cleaned encounters this :;eii()rita 1'a:;s0 11 through a .-chool follo\\ ed this pattern--ob\'iously it i a regular of \'Cry small ( 10 mm) Jllacbmiths, several pattern in the hehavior of the species. of which ho\"ered head-down in its path: ho\\, e\"e r, the cleaner showed no interest in these n:h. ()fh('1' illtem('tioIlR-Ob. ervations are too few On t,"o ncca:;WllS it picked at a piece of drifting to recognize c1i:tinctive a pects in cleaning in dehris. teraction. in\'olving the many other pecies that occa. JOnally are cleaned hy enorita. 'Csually such activity i: noted imply a occasional ight Sf/I()I ita-kd"rish illtl mrli"/ls.-At least one ing.' of .mall clusters of fi . h, or individual fis h, :;pecies regularly lIIitiate:; cleaning bouts with h'I\'ering hefore a senorita. In all uch en ..;enoritas. Earlier (llohson, l~f;:Ja) I reported counter, ho\\'ever, only the external body urface o\)::;eninp: a kelpfbh, lId! J'IIsfidlliS /'(}s/ mtll8, was cleaned. repeatedly soliciting cleaning from unresponsi\"e :\' ote. follo\\' regarding two other species that seii.oritas. The kelptlsh was concealed among are cleaned by .enorita . henthic algae, which is the typical habitat of this A . ingle senorita wa ob en-ed cleaning a pile l1:;h. But each time it senorita approached in perch, after which its activity wa noted for the water o\"erhead, the kelpfish rose up into 15 min. Pile perch are not abundant where these the sefiorita's path, \\'here it hO\ered motionless, ob en'ations were made, and after leaving the fins erect (Fig-me ,)). A succession 'If :;enoritas fir. t individual, the senorita swam alone in mid water for 14 min. It moved over a wide semi circular cour e during thi time and showed no interest in any of the many fi h that it pas ed, although none were pile perch. It did pick at three small item drifting in midwater, After 14 min it made an abrupt course change and, with slightly accelerated swimming, went di rectly to a solitary pile perch that was in mid water about 10 m away. The senorita swam about the pile perch, which now hovered head down, but after a close in pection lasting about 10 sec it moved on without picking at the pile perch's body, Many of the fishes cleaned by senoritas occa sionally start, as if nipped too vigorously. Some
FIGl'RE 5.-Kelpfish ho\'ering in midwater, fins er ect, times such fish dar t away, thus terminating the to solicit cleaning from a passing senorita. cleaning. Other times they actively turn on the
502 IIOBSO N: CLEANI G SYMBI OS IS
senorita and drive it away. The rubberlip perch parasitic copepod.', and isopod.. lIe \\'a' nut was noted taking the latter course of action per more s pecific than this. no!' did hc pre.' Ilt daul. haps more often than the other species, even Yariou .. other cleallcI'. rep()rtedly tak nut only considering the relatively few times it was ob ectoparasites but also dist·a .. ed and 11 'c rotic ti. served being cleaned. s ue (Feder, l%fi, and others.) Although the kelpfish is the only s pecies that To determinc ,iust \\ hat it is that "('ll()rita.' was observed consistently initiating cleaning rcmo\'e from the hodies (If (lther lishe .. I x from senoritas, individuals of sevenl1 species do amined the g ut contents ()f::'7 specimen .. . 111 to so in at least one situation. This occurs where 17:) 111111 long, that \\'ere speared \\ hill· tht'~' \\. l' exceptionally la r ge concentrations of sei"i orilas, cleanin g' other fishes. Food ilt-m' ill t1ll'ir gout ... sometimes thousands, swim abo\'e the r ocks. ranked as percentage of ('ach it'm ill the ntlr' At various times, garibaldis, pil perch, rubber sample, \\·cn.· as follows: cali~dcl copepods. 39', ; lip perch, olive rockfish, and others were ob gnathiid isopod lan'ae, I::" (; algal' with en served hovering in the soliciting posture amid crusting I))'rozoalls, 10' (; caprellid amphipods, these concentrations (Figure 6) until one of the G' , ; Ii"h ;:cales. J (( ; alld frag'IlI('l'h of Illllipar senoritas approached and cleaned them. asitic Cl'ltsLtc ans, l ' ( . l 'nidelltitied material made up :!'(j', uf the sample. Of the ::'7 .peci Material R emoved from Other Fishes by mens, ectoparasite:-. occulTed among the gut Coll the Senorita tents of all Intt i\\'(). JIl n1Ost, tlte pctoparasit·s predominatcd. E\ en t1Hlugh the d.lta an' COll Food habits of clean ing s6ioritas.-An ob\'i \'incing. they do not full~' rcflect the e,tl'llt to ous ques tion is : What do sei"ioriias remo\'e from \\'hich cleaning' oj,\'iousl.\' domiIlated the acti\,lty the bodies of fi s hes they clean? Limbaugh of lhese particular fi,;h for at ]p;tst .. ('\'eral hour,; (1955) stated that senoritas r emO\'e bacteria, leadiIlg' up to their capture. ThIS i.' I,ecallse the
FICl' RF. G.-Garibaldi hO\'l'ring- amid a ian:-" a""'''111. .1'::' of "'1-' nt:l: FISIIf.RY BI LLF.TI:-: \ OL 69. :\0 3 parasites are so small (1-3 mm long) in com chids, which were found on 12 species of the pariso n with the size of other food item. For fi hes sampled, a re mobil e forms about 2 mm example, 471 gnathiid isopod larvae were present long (all lengths of parasites here and below do in one senorita gut, but being so small they con not include egg ca es) that occurred mostly on stituted on ly 35 r ; of the material. On the other the gills of their host . The cali gids, which in hand. a fe'" algal fragments. with encru ting fested 29 species of the fishes, are mobile forms, bryozoans, made up 40r;. of the material in this 2 to 4 mm long, that occurred mostly on the ex same specimen. However, on ly seven individ ternal body surface of their hosts, although two uals contained ectoparasites alone, as compared species were found only in the oral cavity. The \\'ith 17 that contained both parasites and free dichelesthiids are highly modified forms about living prey. In all but one of these, the two :2 mm long that were attached to the gills of two classes of material " 'ere sharply divided in the pecies. The lernaeid i a highly modified form gut. usually " 'ith the free-li\'ing material poster about ;") mm long that wa attached to the fins iorly in a more advanced stage of digestion. Al of 12 species. The chondracanthids, which in though most of the ectoparHsites in the gut had fested five species, are high Iy modified forms, 3 or undergone extensi\'e damage and would not, by 4 mm long, that lived attached in the branchial themselves. ha\'e been identifiable to species, this chamber, including the gills, of their ho t. The material usually gr aded gradually to freshly in lerneopodids, infesting eight of the fish species, gested pecimens that were readily identified. are highly modified forms, 2 to 5 rom long, mostly This fact, coupl ed with the circum. tance that in li\'ing attached in the branchial and oral cavities, di\'idual senoritas tend to stay pretty much with although one individual fi h carried everal at a single type of food organism during a given tached to its dorsal fin. Finally, the argulid period . greatly aided the task of analyzing this i a mobile form about 2 mm long that was found material. on the outer body urface of one species of fish . The fish pecies hosting r epresentatives of the EctOl)(ll'asites 011 the fis hes.-To asse s the sig different copepod and brachiuran families are nificance of cleaning in remo\'ing ectoparasites, Ii ted in Table 2, and example of the six copepod one must know what parasites occur on the fi hes, families are illustrated in Figure 7. as "'ell as the extent of the infe tation. Thus Thus a variety of ectopara itic copepods occur the su]'\,€y of ectoparasites done in conjunction on the fi hes, but only caligid were found among with this "'ork includ ed es entially every fish the gut contents of the cleaner. Further· species exceeding 100 mm long r egularly present more, although 13 pecies of caligid (five spe· in the La Jolla study area, as well as every spe cies of the genus Caligus and eight species of cies that \\'as seen being cleaned there. Ecto the genus Lepeophtheil'us ) occur on fi hes in thiE parasites infesting these fishes incl ude 33 spe area, only a relatively fe'" of these are significant cies of copepods. one species of brachiuran, two as prey of the cleaners, as noted below. species of isopods, one species of leech. and one Of the two isopods, one, Lit'oneca vulga1'is, a species of monogenetic trematode. Following is large parasite, about 20 rom long, was found in a brief summary of the information being com the branchial chamber of j ust one species of fis h piled on these parasi tes in collaboration with R. and was not found to be pr ey of the cleaners. F. Cressey. On the other hand, the highly mobile gnathiid Copepods are the predominant ectoparasites larvae (Figure 8) , which are about 2 rom long, on fi shes in this ar ea. The 33 species represent ar e a major prey of the cleaners. Only one form seven families: Bomolochidae (6 species) , of gnathiid was r eadily recognized, but mOrE Caligidae (13 species), Dichelesthiidae (2 spe than one species may occur among this material. cies), Lernaeidae (1 species ), Chondracanthidae Par asites of the body surface of fi shes, the gnath· (5 species ) , and Lerneopodidae (6 species) . The iid larvae were taken on 11 of the fi sh s p ecie ~ one species of the closely related brachiurans is sampled, but I suspect that they ar e actually a member of the family Argulidae. The bomolo- more widespread and abundant than these data
504 HOBSO : CLt:AN ING SYMBIOSIS
TABLE 2.-The types of ectoparasites and the fi shes they infest, based on a survey of the fi shes in the study area. Wher e more than one species of parasites is included under a heading, the number in parentheses following the name of each fi sh thereunder indicates how many different species of that type a re represented on that fi sh. For a ll fi sh species li sted, t he number of infested inJividuals is shown over the number of individua ls ex amined, followed by the r ange in numbers of individual parasites of that type which were taken from that species of fi sh .
COPEPODS COPEPODS-Cont. ISOPO DS BOMOLOCHI DAE (6 species of 3 genera ) CALI GIDAE-Cont. CY MOTHO I DAE (I speci es) Fi.shes infested w ith membe rs of .Jlherinap, a!fi ni, I I ) 13/ 13, 1·23 Fis h infested with this parasite: this fami ly : Pll'uro n irn cnys COr' n OHIJ (I) Si l O: 1-5 SeboJlts m yst inus 1/ 9 : 1 Paralab rax dalhrat u, ( I ) 2/ S , 1·2 (2) 5/ 11 , 1-20 DICHELE STHI IDAE (2 spe cies of I genus) P. ntbuli/tr GN ATH II D LARVAE (number of species not Phantrodon alripfJ (I ) 1/ 13: I Fishes infested with parasites determined) of this fami ly, R harachilu, va rca (I) 3/ 15, I Fishes infested with these para si tes: C;ymTlotnorax morJax (1) 1/ 1: 9 lUicromdrus mi nim us (1) 1/ 7: 1 Chromis punclipinnis 1/ 10: I Parafabrax "tbullia (1) 5 / 11: 2-70 HYPJy popJ rubicllllda (1) 1/ 20: 1 I/ yps~.. pops rubicu nlia 4/ 20: 1-8 Srarpac na gllllaia (I ) 6/ 14 , 2-56 LER N AEIDA E (I species) Pi mt/omtlopon pulchru m 3/ 14 : 1-4 SrboJln mystinllJ ( I ) 1/ 9: I Fi shes infested by this parasite: S"basi n alrotJirtTls 4 / 16: 1-20 S. u rranaid" (I ) 2/ 11 , I-S TrarhllrllJ s\' m m(tn(lU (1) 1/ 7: S. (arnat us 3/ 11 : 1 ux),"biu, pic l u, ( I ) I / S, I .1"iJot rt' m Il J lio,".. ,idJoni (1) 1/ 8: I S. (hr"JOmtias 1/ 11 : 2 A,h"inap, a!fi ni, (I) 2/ 13, 1-4 Clttl/otrema salllr flum ( I) 2/ 16: 1-2 S. rOTlJldlallu 1/2: 1 P/t uronirhthys (0010111] (1) 6/ 10: 1- ' 1 .\ll·dw/una (ol1lor"i,.,1111 (J) 5/ 13: 2·7 S. sura noilitJ 1/ 11 : 1 Bra( ltyisllUs Irt'natus (I) 1/ 5: 1 CALIGIDAE (13 species of 2 genera ) S urrirl'ps 3/15: 1-5 Emhiotora JtlC/.:JOfl l (I) 2/ 15 : 1-2 O.\y/ebius PICtu S 3/ 8: 1-3 Fishes infested with me mbe rs o f this family; IIvpJl/rus rflryi (I) 2/ 11: 1·2 Scorpaolichthys mar m ora/ us 3/ 10 : 2-5 Pora/abrax d otArat ui (2) 4/ 8: 1-5 /'ltallfrodoll jurea/us (1) 3/ 12: 1·3 P. n,buli/" (3) 4/1 1, 5-27 k ll(lroehllus toxo/tJ ( 1) 1/ 10: MONOGEN ETIC TREM ATOD E (I species) Ca ula/ali/us prinu pJ ( I ) 1/ 4: R ,'arca (I) 2/ 15 , I Fishes infested w ith this para si te: J1 n 1J olrrmUI davidJo ni ( 1) 1/ 8: 1 J/ l(romdrlls 11/I /l l1nu s ( I) 5/ 7: 1-6 Alt'dialll na co /ilornlt'nJis 61 13 : 2-16 Chriiotrrma sa turnum (1) 2/ 16: 1 :/ thrronaP' a/fin" (I) 2/13, 1-4 Glrdla nij; rira ns 4 / 10: 1·4 RltarocltiiltJ t oxott'J 2/ 10 : 1 Afrdialu na californiensis (2) 12 / 13: 1-75 CHONDRACANTHIDAE (5 sp.coes of 5 genera) R .'a((o 2/ 15 : 1-18 Gird/a nigrjco nJ (3) 1/ 10: 1- 14 Fishes infested wi th parasites of E mbia' ara lack'ani (2) 2/ 15 , 1·2 this family: lfyps'YPops ruhu unda 1/20: 1 Phantrado n alrip" (I) 5/ 13, 1·5 OXYJIl/is eail/ornl(o (1) 13138 1-4 Pl rn tiomctopoTl pulchrum 10/ 14 : 1-26 Rhacochilu, laxal" (3) 10/ 10, 1-10 Srorptlt'fla gllrt ala (I) 2/14: 1-5 S(orpacna f,lIttllta 2/ 14· 5-8 R . va« a (2) 4/ 15 , 1·11 S(orpatnichth ys marmoralllJ {Il 4 / 10: 1-4 StbaJto alrO:lft"ns 7/ 16: 1 Chro rn iJ p ll Tl ctipinniJ ( I ) 10/ 10: 2-39 lIt'iaosticltus rostratuJ (I) 7/ 13 : 1·7 S. rOllllrlltitus 1/2: 1 Hyp,), PDP' ru bicunda (2) 19/20 , 1· 144 P/tllronl(htItYJ ronlOill! ( 1) 7/10: 1·47 S. minitlt li S 2/5: 1·2 S. u rranoidt.'s 3/ 11 : 1-26 Pirn t /omttopon pulc hrum (3) 13/ 14: 1-70 l ER NE O PODIDAE (6 species of 4 genera ) S. urriceps 2/ 15 : 1·3 OX), luli, cali/arnica (3) 13 /3S, 1·59 Fishes hfested by porasites of jfeteros ticJII4 J ros/ratU J 6/ 13: 1·7 Scorpotno guttato ( I) 7/ 14: 1-14 this fami ly: StboJttJ atrotliffnJ (3) 6/ 16: Cltri/otremo wturflum (1) 1/ 16: LE[C H (I species) S. carna'u, (2) 3/ 11 , I (;,,"Ia nigro 505 FISHE RY B\,; LL ETI : \·0 L. 69. NO. 3 parasites are 0 small (1-3 mm long ) in com chids, which were found on 12 species of the parison with the size of other food items. For fishes sampled, are mobil e forms about 2 mm example, 471 gnathiid isopod larvae were present long (all lengths of parasites here and below do in one sei'i orita gut . but being so small they con not include egg cases) that occurred mostly on stituted on ly 35(" ( of the material. On the other the gills of their hosts. The cali gids, which in hand. a fe\\' a lgal fragn1ents, with encrusting fested 29 species of the fis hes, are mobile forms, bryozoans, made up 40(( of the material in this 2 to 4 mm long, that occulTed mostly on t he ex same specimen. However, only seven individ ternal body surface of their hosts, a lthough two uals contained ectoparasites alone, as compared species II'ere found only in the ora l cavity. The with 17 that contained both parasites and free dichelesthiids are highl y modified forms about living prey. In a ll but one of these, the two 2 mm long that were attached to the gills of two cl asses of material \\'ere shar ply di vided in the species. The lernaeid is a highly modified form gut. usually with the f ree-li\'ing material poster about 5 mm long that was attached to t he fins iorly in a mor e advanced stage of di gestion. Al of 12 species. The chondracanthids, which in though most of the ectoparasites in the gut had fested fiv e species, are highly modified forms, 3 or undergone extensive damage and would not, by 4 mm lon g, that lived attached in the branchial themselves, ha\'e been identifiable to species, this ch amber , including the gills, of their hosts. The material usua ll y gr aded gr aduall y to freshly in lerneopodids, infesting eight of the fi sh species, gested specimens that were readily identified. are hi ghly modified forms, 2 to 5 mm long, mostly This fact, coupled with the circumstance that in living attached in the br anchial and oral cavities, di\'idual enor itas tend to stay pretty much with although one individual fi h carried several at a single type of foo d organism dming a given tached to its dorsal fin. Finally, the argulid period, greatly a id ed the task of analyzing this is a mobile form about 2 mm long that was found material. on the outer body surface of one species of fish. The fi sh species hosting representatives of the Ectoparasiteli 0 11 the /is hes.-T o assess the sig different copepod and brachiuran families are nifica nce of cl eaning in r emov ing ectoparasites, li sted in Table 2, and examples of the six copepod one must know what parasites occur on the fi shes, families are illust rated in Figure 7. as " 'ell as the extent of the infestation. Thus Thus a variety of ectoparasitic copepods occur the sun'ey of ectoparasites done in conjunction on the fishes, but only caligids were found among with this work included essentially every fi sh the gut contents of the cleaners. Further species exceeding 100 mm long regularly present more, although 13 species of caligids (five spe in t he La Jolla study ar ea, as well as every spe cies of the genus Caligus and eight species of cies that was seen being cleaned there. Ecto the genus L e7Je ophtheirns ) occur on fishes in this parasites infesting these fi shes include 33 spe area, only a relatively few of these are significant cies of co pepods, one species of brachiuran, two as prey of the cleaners, as noted below. species of isopods, one species of leech, and one Of the two isopods, one, L ivonecct vulga1'is, a species of monogenetic t rematode. F ollowing is large parasite, about 20 mm long, was found in a brief ummary of the information being com the branchial chamber of just one species of fish ])iled on these parasites in coll aboration with R. and was not found to be prey of the cleaners. F. Cressey. On the other hand, the highly mobile gnathiid Copepods are the predominant ectoparasites larvae (Figure 8 ) , which are about 2 mm long, on fishes in this area. The 33 speci es r epresent are a major prey of the cleaners. Only one form se\'en famili es: Bomolochidae (6 species) , of gnathiid was readily recognized, but more C11.ligidae (1 3 speci es) , Dichelesthiidae (2 spe than one species may occur among this material. cies). Lernaeidae (1 species) , Chondracanthidae Par asites of the body surface of fishes, the gnath (:'J species) . and Lerneopodidae (6 species ). The iid larvae were taken on 11 of the fi sh species one . pecies of the closely related brachiurans is sampled, but I suspect that they are actually am mber of the family Argulidae. The bomolo- more widespread and abundant than these data 50 1 HOBSON : CLEANI NG SYMBIOS IS TABLE 2.-The t ypes of ectoparasites and the fishes they infest, based on a survey of the fi shes in the study area. Where more than one species of pa rasites is incl uded under a heading, the number in parentheses f oll owi ng the name of each fi sh thereunder indicates how many differ ent species of that type are represented on that fi sh. For a ll fi sh species listed, the number of infested inJividua ls is shown over the number of ind ividuals ex amined, followed by the r ange in number s of individua l pa rasites of tha t type which were taken from that species of fi sh. COPEPO DS COPEPODS-Cont. ISOPODS BOMOLOCHIDA E (6 species 01 3 genere) CA LI GIDAE-Cont. CY MOTH OIDAE (I species ) Fish es in fested w ith members o f .1 Ihainop, a[fi ni, (I) 13/ 13, 1-23 Fis h infested w ith this para site : this family: Pfl' uronicn tn\'J (DOl/HilS (1) 5/ 10: 1-5 SebaJ!n mYJ tinuJ 1/ 9: 1 Pa, alab,ox c1alh,alu, (1 ) 2/8, 1-2 DICHE LESTH IIDAE (2 species 01 I genus) P . ntb uli/u (2) 5/ 1 I, 1-20 GNATH II D LAR VAE (number of species not Phant rodo n atripn (1) 1/ 13: Fishes in fested wi th parasites determined ) of this fami ly: Rha,o, hil u, vo((a (I ) 3/ 15, I Fishes infested wi th these para si tes: G\'m nothorax morda.\ (1) 1/ 1· 9 A/icr omt lru J minimuJ ( I) 1/ 7: 1 ChromiJ punrlipinn lJ 1/ 10 : 1 Paralabrax IItbllfi/a (1) 5/ 11 : 2-70 H y p'ypop' , ubi"",da ( I ) I 120, I l/yprYPopJ rllbicu nda 4/ 20: 1-8 S(orpotna gut/a la (1) 6/ 14 : 2-56 LER NA EIDA E (I species) Pimdomdopo n pulchru m 3/14 : 1-4 Srbasi t J m yst inuJ (1) 1/ 9 : 1 Fis hes in fested by this parasite: StbaJtn atrovirt nJ 4/ 16: 1·20 S. , m o noid" ( I ) 2/11, 1-8 T rachllrlii J\,mmtlr/(III (I ) 1/ 7: S. car nal ll J 3/ 11 : 1 Uxy/tbiuJ pict Ii ] ( I ) 1/ 8: I ,lniJo/rf' m uJ da:·idJoni {Il 1/8: 1 S. cn rYJo mlofaJ 1/ 11: 2 Alhuinop, a{fi ni, (I ) 2/ 13, 1-4 Cnri/otrtma J(lturllu m ( 1) 2/16: 1-2 S. co nJteff atu J 1/ 2: 1 P/ru ronichtnYJ (Ol' n OJIl! (1 ) 6/ 10 ; 1-1 1 J /(diafu na cali/orllil·nJIJ ( I ) 5/13: 2-7 S. u rra noidn Ji l l : 1 B rach),iJlillJ frlona/IlJ ( I ) l i S: 1 S. u rriCt'pJ 3/15: 1-5 CALIGIDA E (13 species 01 2 genero ) Emhi%co }(lebOlll (1) 2/15: 1-2 O.n·fc'biuJ pletlii 3/8: 1-3 Fishe s infested with membe rs of this family: If,,pJllruJ cnryi (1) 2/ 11 : 1·2 ScorpocniehtnYJ mar m ora/ uJ 3/ 10 : 2-5 Paralabrar datnratuI (2) 4/ 8: 1-5 Pha 1Z frodoll /urc(ltllJ ( I ) 3/ 12 : 1·3 P . n, b uli/ u (3) 4/ 11 , 5-27 k hacocJllluJ tOolol(J (I) 1/ 10: MONOG EN ETIC TRE MATODE ( I spec ies) Ca u/olati/u! pri R.apJ ( 1) 1/ 4 : R ,'a ((a (I) 2/15 I Fishes infested wi th this parasi te: A n1J otrt muJ da vidso ni ( ' ) 1/ 8: I .1/lfromt'lrllJ rnllllrnlli ( I ) 5/ 7: 1·6 A/tdialu na c(1. /i/ ornll' nJiJ 61 13 : 2-16 Cluiiotrr ma soturnu m ( 1) 2/ 16: 1 .1thalnopJ affjnlJ (1) 2/13: 1-4 Glrel/a nigrica nJ 4/ 10: 1-4 k hacocn iiuJ toxotn 2/ 10 : 1 t\ftdialu na (ali/ornlt RJiJ (2) 12 / 13: 1-75 CHONDRACANTHIDAE (5 speGes 01 5 genero) R "0((0 2/ 15 , 1-18 C irtlla n igrico nJ (3) 1/ 10 : 1·1 4 Fishes infested wi th parasi tes of E m biolo,a Ja ' k ,oni (2) 2/ 15 , 1-2 this fami ly: l/ypn'popJ rl4 bicunda 1/20: Pimrio mctopon pulchrum 10/14 . 1-26 Phanuodon al, ip" ( I ) 5/ 13, 1-5 OX~."Ju / 1J cailjornlC(1 ( 1) 13/38: 1-4 Rha,o,h ifu, l ox ol" (3) 10/ I 0, I -10 Scorparno f;uttata (1) 2/14· 1-5 Scorpot' na gullata 2/14. 5·8 SrbaJto atrO:'lr, nr 7/16: I R . va((o (2) 4/ 15, I -II ScorpatnirhthYJ marmora/III ( 1) 4/10: 1-4 S. COlil/fl/a/UJ 1/2: 1 ChromiJ pllnclipinniJ (1) 10/ 10: 2-39 lft' taoJtuhllJ rOJ/ ra/IIJ (1) 7/ 13: 1-7 S. minw lu J 2/5: 1-2 H )·p,ypop, ,ubi, u nd. (2) 19120, 1-1 44 Pleuron lchth-;"J COt/IQJIIJ (1) 7/10: 1-47 S u rra 1l oldo 3/ 11 : 1-26 Pimd omttopon p ulchrum (3 ) 13/14: 1-70 lERNEOPODI DA E (6 species of 4 genera) S. u rricrpJ 2/ 15 : 1-3 Oxyjuli! cali/o,ni, a (3) 13/ 38 , 1·59 Fishes i"fested by poro si tes 01 I/t t froJ tiehuJ rOJ trat UJ 6/ 13: 1-7 Scorpal' na gut lala (1) 7/ 14 : 1-14 this fam il y: StbaJtn atrovirt nJ (3) 6/ 16: Cht'iIQl rt' ma Jat url/llm (I) 1/ 16: 1 LEECH (I species) S. ,a'nalu, (2) 3/ 1 I, I (; " ,1/. (I) 1/10, I n i~,, 505 FlSIl ERY Bt:LLETJN : VOL. 69, NO. J A c D E F FIGl' RE 7.-Representatives of the families of ectoparasitic copepods f ound to inf est fi shes in the La Jolla a rea. A. Bomoloch idae (Bolllo1oclws /ongicandll.s, female, after Cressey, 1969b) ; B. Caligidae (Caligus hobsoni, male, after Cressey, 1969a) ; C. Dichelesthiidae (H atschckia pacifica, female, after Cressey, 1970); D. Lernaeidae (Penicllins fis sipes, fema le, after Wilson, 1917) ; E. Chondracanthidae (Chondmcanthus g?'acilis, female, modified a fter Wilson, 1935); F . Lerneo podidae (Epib1'anchiella septicauda, female, after Shiino, 1956). 506 HOBSON : CLEANI NG SYMBIOSIS topsmelt each carried from 1 to 23 specimens of the copepod Cal'igus sermtus on their body surface. Two topsmelt also carried the copepod Parabomolochus const1'ic tus on their gills, a single parasite on one, four on the other. Two topsmelt also carried the copepod Peniculus fissipes embedded in their fins. Six of the eight senoritas that had been cleaning topsmelt when collected had ectoparasites among their gut con 1 mm tents. Five contained only Caligns se1Tatus as many as 73 in each fi h. One other contained only 10 gnathiid larvae, a para ite that was not FIGURE 8.-Gnathiid larva from the body surface of the black-and-yellow rockfish, Sebastes chrysomelas. seen on t he topsmelt themselves; however, as noted above, I suspect that this parasite is more widespread than our survey data indicate. fis hes that rarely or never interact with cleaners. Six species of ectoparasites were collected In considering the ectoparasites found in the gut from 20 garibalclis, 18-1 to 240 mm long. ine of particular senoritas, it would be most mean teen garibaldis each calTied 1 to 144 Caligns hob ingful to do so in regard to the ectoparasites sOlli on their body 'urface. Thirteen each carried known to be hosted by the species of fishes that 1 to 4 individuals of an unidentified species of these particular senoritas were cleaning when L epeophtheirns on their body surface, and one collected. Of the 27 cleaning senoritas taken carried a si ngle Eomolochlls ardeole in its for the gut-content analysis, 15 (561" ( ) were branchial cavity. In addition, four carried 1 cleaning black miths, 8 (301"() were cleaning to 8 gnathiid isopod larvae, two carried a single topsmelt,2 (7 r;t ) were cleaning garibaldis, and leech, and one carrif'rl a single monogenetic 2 (7 c ) were cleaning halfmoons. Thus the trematode, all on thei I' body surface. The two selection closely parallels the relative frequency senoritas that were col lected as they cleaned with which senoritas were observed cleaning garibaldis had preyed mostly on gnathiid larvae, these same species (Table 1) and is a good sample with each containing over 400 of these parasites. of the fishes that are cleaned by senoritas. In addition, one had consumed six CaliglIs hob Three species of ectoparasites were collected SOl/i. and the other had taken five Lelleo]Jhtheims from 10 blacksmiths, 141 to 199 mm long. Each sp. of these blacksmiths carried from 2 to 39 indi Four species of ectoparasites were co llected viduals of the copepod Caligus hobsoni on their from 13 halfmoons, 166 to 295 mm long. Twelve body surface. One specimen also carried a single of the 13 halfmoons each carried 1 to 75 Caligns gnathiid isopod larva on its body surface, and hobsol/i on their body surface. Each of two also another the copepod Clavello7Jsis flexicunicn on carried a single L epeophtheil'Us sp. on its body a gill arch. All 15 senoritas that were collected surface, and each of six carried 2 to 7 PeniCIllus as they cleaned blacksmiths contained either fis::;i lJ eS embedded in its fin s. In addition, each Caligus hobsoni or gnathiid larvae, but no other of six carried from 2 to 16 monogenetic tt'ema ectoparasites: one contained gnathiids alone, todes on its body surface. Of the two senoritas seven contained C. hobsoni alone, and seven con collected as they cleaned halfmoons, each con tained both gnathiids and C. hobsoni. Up to 256 tained only Caligus hobsoni in its gut contents, individuals of C. hobsoni and up to 263 gnathiid one a single specimen and the other eight. larvae were counted from among the stomach Significantly, with the exception of the gnath contents of individual senoritas that had been iid larvae in a cleaner of topsmelt, as discussed cleaning blacksmiths. above, no parasite was found in the cleaner's Three species of ectoparasites were col lected gut contents that did not occur on the species f rom 13 topsmelt, 122 to 212 mm long. These of fi sh that was being cleaned by the cleaner 507 F ISHERY BULLET IN: VOL, 69. NO, j \\'hen it wa, collected, Thi fact f urther sup t he body surface of a single sheephead, and t hi po rts the contention that cleaning tends to be fis h has not yet been seen being cl eaned, F ur species-,'pecific for a gi,'en senorita, t hermore, up to 4 gnathiid lar vae, which cl eaners The data clearly show that the para ites most take f r om other fi sh, were f ound on 3 of the frequentl~' taken h~' se i1 0ritas a re certain mobile sheephead, Similarly, the treefish, Sebastes se1' forms that occur on the body urface of t heir I'ic eps. which i heavil y infested wit h caligid , host. It may be that other para ites on the ex has not yet been seen being cleaned. The tree ternal body surface are not taken, "t\o leeches or fi sh is not known to calTY C. hobson i, but 13 of t renntories \\"ere found among- g- ut contents. F i pecirllens examined carried up to 12 L epeoph ,'en thoug-h these forms are abundant on th eil'll.') 10llgi/les on their body surface, and 3 the garibaldi and halfmoo n, Al so. the guL con carried up to;- gnathiid lar vae, T he significance tents did not sho\\' evidence of the lel'naeid of these exceptions to what seems a vali d gener P (,I/indlls jiss/ll(,,~ . an immollile form \\'hich par a li zati on has not been determined, Perhaps it tially embeds itself in the skin of its hosts is sig' ni ficant that t hese two specie of fi sh ar e mostly on the Ans, This parasite occurs on top not hem'ily infested by copepods of the genu smelt. garibaldis, and halfmoons among those Caliglls. as are the more frequen tly cl eaned kno\\'n to be cleaned hy senoritas, Howe"el'. fi shes. negative e\'idence based on the meager gut-con The many parasites that infest the or al and tent data are weak, especially as the clean in g branchi a l cavit ies might seem to be potential lallrid Lah roides IlhthirolilwlIlis in IIawaii , prey fOl' cl eaners, but I found no evidence that \\'hich feeds mostly on calig-o irl cO]Jepods, f re the, e parasites ar e taken by senoritas, quentl~' takes lel'naeid s (Randall. 19:58; Young , T he principal ectopa rasite on the body sur bluth, 1968). I would expect additional study face of the two most f requent ly cleaned fi shes, to sho\\' that cleaning' seiioritas at least occasion the blacksmith and t he topsmelt, ar e the copepods ally take P. ,{iss illes , )J e\'edheless, e,'el'a l Ca!i{Jlls hobsoni and C, Sel' l'atus, respectively, abundant ushes infested by p, /i.-.;.-.;illes, but not \\'hi ch a re very imilar to one another mor pho found to carry calig- id s. gnathiids. 01' other mo logicall y, Wit h just one exception among the hi Ie external forms. were not seen bei ng cleaned, fi shes surveyed (discussed below) , C. se1'l'atus For example. the white sea perch is one of t he seems to be restricted to topsmelt. On the other most ahund,tnt species at the 3- to 10-m station ha nd. C. hobsoni occurs on a wide variety of (Iff La ,Jolla and ~'et \\'as ne,'er seen being cleaned, speci e a nd is also the pri ncipal form on gari T\\'el\'e specimens of this fish ,,'ere examined, baldis a nd halfmoons, Interesti ngly, a list of and the onl~' ectoparasites found \\'ere one to t he fi she hostin g this pa rasite, r anked by in f,.ul' p, (i'-';'-';//I('-'; on th ree inrli"iduals, Simi larly. ci dence (Table 3 ) , looks m uch like the ranking the onl~' parasite found on 11 rainbow seaperch, of fi shes that were obser ved being cleaned by the an ahundant species in the study areas that was senorita (Tab!e 1). not seen heing cleaned. was a single P. /issiJles .. n one inrli"idual and t\\ 0 on another. Th e importance of cleaning in 1'e ducing t he Ho\\'e\'er . not all flshes whose external body incidence of ectopal'asit es on /ishes,-Certainly ,urLtces are hea\'il,\' infested by mohi le for m cleaners r emove many ectopar asites f rom the \\ 1']"(' ohsl'ned heing- cleaned, The sheephead. bodies of cer tain fi shes-t he numbers in their l'lt! 11"11/11"/,1111 /, I,'!Jo/Jll. is a case in point. diet altest to thi fact. But does cl eani ng in ('I/III/lIS h,,/,'-';III/i occurs on this ush, hut only in fact appreciably r educe the level of infe tation fr(>'illl'nll~' a single specimen of this copepod on the e fi hes, 0 1' do other para ites qui ckl y \\ a ... takt'n from eath of ~ of the 1 -1 sheepheads replace those that are removed by the cleaners ? tlla \I l're L'xaminer!' llowe\' r. the sheephea 1 AILhough this que tion is difficul t to an weI', some i twa' ih' infl'." cd h~' two ,]I tie. of Lc 11(,()/IIz/h ei in , ight i, ]wo"ided by obser vati011 s on the gari I'll • • a }.!l'nu..; of cop pod::; that is c10s Iy l' lated baldi, When guarding eggs on their nests dur til ('filii/III', t pto 10 L, IJ(I/TIIS were tak n from ing th l' productiv ea on, male gariba ldi be- 50 I-IOB50 : CLEANING SYMBIOSIS come especially intolerant of the presence of fested fi shes. The samples included too few of other fish species. Clarke (1970) recorded the the other parasites to make a meaningful com number of times garibaldis, in defence of their parison. It remains a question why Lepeoph territory, attacked fi sh of various other species theil"us sp. and the gnathii d larvae did not show at different times of the year. He found that a pattern of occurrence similar to that of C. when males wer e guarding eggs their attacks on hobsoni, as both of these parasites are known enoritas increa ed elevenfold. Not surpris to be prey of the cleaners. In any event, these ingly, I saw no cl eaning of garibaldis that were data add to the evidence which indicates that guarding eggs. At other times of the year male C. hobsolli is the primary prey of cleaning sen garibaldi do not guard their territory as vigor oritas in the study areas. ou Iy against members of other species and are frequently seen being cleaned. A series of these Ectoparasi res on Senori ras males were coll ected both in and out of the re productive season, and the numbers of ectopar Senoritas that were cl osely observed as they asites they carried were assessed. Seven indi cleaned other fi shes often were noted to have viduals (mean length 228 mm) sampled as they cali gid copepods on thei r bod ies. On e senorita, guarded their eggs carried a mean of 67 Ca/iglls about 120 mm long, was host to an estimated hobsoni (range 20-144),2.5 L epeophtheinu; sp., 100 of these parasites concentrated especially 1.4 gnathiid isopod larvae, and 0.2 monogenetic along the dorsal-fin base. These observations trematodes. These counts contrast strikingly were significant because during the survey for with those from six males (mean length 219 mm) ectoparasites, most senoritas taken from the pop sampled outside the reproductive season, which ulation at large were free of external forms, carried a mean of onl y 4.8 C. hobsoni (range although many carried a chondracanthid cope 0- 13 ) , 1 L epeophtheims sp., 0.8 gnathiid larvae, pod on their gills. and no monogenetic trematodes. These findings Twenty senoritas, 10 ~ to 190 mm long, were suggest that males which are guarding eggs be sampled from among those giving no indication come heavily infested with C. hobsoni when they of being cleaners. Eight of these carried 1 or do not allow cleaners to approach them, a con more of the chondr acanthids on thei '· gills, but clusion strengthened by the fact that 0\ er this only 2, or 10 ((, had parasites on their external same period the relative numbers of this same body surfaces: one of these carried 10 speci parasite were not noted to change on other in- mens of Ca ligus hobso lli and 1 specimen of TABLE 3.- Hosts of Caligus hobsoni. Number o f (:. habJoni Percent Specimens Specimens Species hos tin g on each occur· examined C. nobsQni infested fish renee mea n (range) Blacksmith. Chr omlJ pu nct lplnnss 10 10 10.6(2·39) 100 Topsmelt, A tlllrin opJ affi niJ 13 '13 '10(3·23) '1 00 Garibaldi, H)'pJYPopr rubicu Tlda 20 19 31.2(1·144) 95 Halfmoon, J\ftdialu na colijornit nJiJ 13 12 19 .8( 1·75) 92 Opaleye. C irdla nigriconJ 10 8 54(1· 14 ) 80 Olive rockfish. S,baJtn u rr anoidtJ II 5 2.6(1-4) 45 Blue roc kfish , SthoJtn mYJ tinuJ 9 4 I 44 Shorpnose seoperch, Pho nuodon olriprJ 13 5 2(1 ·5) 38 Senorita , OxyjuliJ roliiorn iro 36 9 11(1 ·59) 24 Sheepheod, Pimtlomttopo n pulrhrum 14 2 I 14 Rubberl ip perch , Rh ororh i/u J toxotn 10 I 10 Cabezan, Sro rpotnirhthYJ marmorotUJ 10 10 Gopher rockfi sh , StboJttJ ror notuJ II 9 Pile perch, Rh ororhil uJ va((o 15 7 Kelp roc kfi sh, SthtlJln olro,,-, irtnJ 16 6 1 :tthuinopJ a{fi niJ does not corry C nobJoni, but rather is the sole host (with one exception, see text ) of the ve ry simi lar C. u rrotuJ. 509 ~b"LRY III LLI,TI ': \'OL 6'1,1>;0 . } LI'JI('()llhl1l ('il'IIS sp.: the other seiiorita carriNI i' 'c d with ['\1' ). s ,norit. Tempcmtllre.-As noted above, the numbers FIG1'RE 9.-Number of senorita cleaning bouts seen dur of seiioritas present at the 20- to 25-m station ing each of 33 observation periods, 15-25 min long, at fluctuated in an apparent r esponse to water different water temperatures in an area 25 m deep at temperature, with the critical level at about 12° La J olla. Periods during which temperature fluctuated to 13° C. Less cleaning occulTed at lower tem were not considered. 11 = number of observation periods at that temperature; where n > 1, value given is the peratmes (Figure 9), which would be ex- mean. 510 HOBSON: CLEA ' ING SYMBIOSIS CLEANING ACTIVITY OF THE (Table 4) . Despite the fact that senoritas were SHARPNOSE SEA PERCH observed to be far more numerous than perch throughout the depth range of this study Unlike senoritas, which clean as adults as well (3-50 m), seaperch performed almost all as juveniles, all of the sharpnose seaperch that I the cleaning observed at the 30- to 35-m lo observed cleaning were juveniles less than about cation, where cleaning by the much more 125 mm long. Occasionally noncleaning sea abundant senorita was limited to a few iso lated perch swim in groups of 15 or more, but those instances. seen cleaning were a lways solitary, or in groups A measure of the incidence of cleaning indi of two or three. In agreement with senoritas, viduals within the population of juvenile sharp cleaning sea perch do not establish well-defined nose seapel·ch was obtained during 39 observa cleaning stations, but instead may clean other tion periods at the 20- to 25-m and 30- to 35-m fi sh at any point as they move from place to locations at La J o11a. These observations, to place. I found no evidence that fi shes which taling more than 26 hI', were made from Sep are residents of other areas come to where sea tember 1968 to February 1969. During this perch are located for cleaning; rather, cleaning period, 201 juvenile sea perch were seen, of which seaperch occur where resident fishes are nu 105, over 52("(, were cleaning other fi shes. Thus merous. As is tr ue of senoritas, E:ea perch use it appears that at least most sharpnose seaperch the same picking technique to clean material are cleaners when they are juveniles, whereas from the bodies of other fish that they use to only a small minority of the sei'iorita population take small organisms from a benthic substrate. seem to be cleaners. Clearly bottom-picking can be preadaptive to cleaning. Cleaning by seaperch, as by senoritas. usuall y occurs within 3 m of the substrate. How Fishes Cleaned by the Sharpnose Seaperch ever, there is li ttle overlap in the cleaning areas Because sharpnose sea perch were observed of the two species: generally seaperch clean at only at depths below 20 m, substantially less data greater depths and/ or in colder water than are available on their cleaning activity than on seno ri tas, where limi ted observations incli that of senoritas. Of the 105 seaperch observed cate they may predominate as cleane,-", e\'en cleaning during the 39 observation periods re when senoritas are m01·e abundant. Data ported above, all but one -we re cleaning black illustrating this di stribution of cleaning acti\· smiths; the lon e exception was cleaning a soli ity at a point in time were obtained at the tary blue rockfish, SebcLstes rnystinns. On two 20- to 25-m and 30- to 35-m locations off other occasions, I saw sharpnose seaperch clean La J olla, where the two s pecies co-occur ing rubberlip perch, but otherwise the only fish seen being cleaned have been blacksmiths (Fig ure 10). Undoubtedly additional observations, TABLE 4.-Number of bouts in which sharpnose seaper ch and seiioritas, r espectively, were seen cleaning other especially in other areas, would expand this list. fi shes during 15-min observation periods at the 20- to I observed sei'iol·itas cl eaning in many different 25-m and 30- to 35-m locations off La J olla. Two obser areas, but my observations of cleaning seaperch vation periods, one at each location, and never more than are limited to La Jolla. Clarke et al. (1967) 45 min apart, were made on each of the dates indicated. saw a sharpnose sea perch cleaning a rockfish at Number o f cl eaning bouts observed 150 m off La Jolla, and Gotshall (1967) reported Do te 30- to 35-m location what he believed to be this species cleaning Mola Seaperch Seoperch Senorita mola off Monterey. Yet no matter how many 22 Nov. 2 13 17 0 27 Nov_ O 7 1 0 different species the seaperch may in fact clean, 9 Jon. 2 5 9 0 there seems no doubt that blacksmiths are prime 15 Jon_ 2 4 8 0 3 ' eb _ 2 12 9 2 recipients in southern California, at least in Tota l 8 41 44 2 depths shallower than 35 m. 511 FISHERY BULLETI ': \ 'OL. 69, NO.3 to have been cleaning blacksmiths, were each kept under surveillance for 15 min, while their activity was monitored on tape, Both swam on il'l'egular course among the rocks but remained 2 within an area encompassing about 15 to 20 m • During this time one entered into 4, the other 5, separate cl eaning bouts, averaging 2.6 (range 0.5-7.5) and 1.8 (range 0,75-3.5) min long, re spectively, all with blacksmiths. The cleaner ini tiated the activity in each instance, but immedi ately thereafter a n umber of other blacksmiths converged on the spot. Most of the cleaning bouts continued after the original blacksmith had left the group, and a succession of others arrived and departed before the bout ended, Although usually they hovered head-down before the clean ers. the black miths nevertheless assumed a wide variety of attitudes. During much of the time they swam \\'ith the blacksmiths, the two sea perch under surveillance closely inspected the blacksmith's bo dies and actually picked at them 18 and 14 ti mes, respectively, Most of the clean in g was directed at the fin bases, particularly the caudal. While in company with the black smiths, one of the seaperch broke away from the grou] and swam to look cl osely at the dorsal fin of a blue rockfish, However, no cleaning oc FIGI'HE lO,-Sharpnose seaper ch inspecting a blacksmith, CUlTed: the blue r ockfi sh swam away as though ",hich ho\'crs to soli cit cleaning, uninterested in cleaning and the seaper ch re tUl'l1ed to the blacksmiths. When not in company Specific Cleaning Interactions - Seap erch with the blacksmiths, the two seaperch swam Bl acksmith alone 1 01' 2 m over the substrate, One descended to the bottom twice and picked at gorgonians: The limited obsen'ations on cleaning by sharp five times on the first descent, once on the nose seaperch pro\'ide details onl y on interac second, tions with blacksmith, As nearly as could be Once a blacksmith was seen obviously attempt seen, when sharpnose seaperch clean blacksmiths ing to present its caudal fin to a seaperch, with the acti\'ity proceeds much as it does when black out succe s in enticing the seaperch to clean, smiths are cleaned h~- senoritas, as described Close inspection did not reveal parasites, but alH)\p, II()\\'e\'er, the o\)sen 'ations were too few par t of the fin was torn away and shredded t!l determine \\'hether or not cleaning activity flesh was exposed. Apparently this blacksmith is c()nsistentl~- initiated by the cleaner. Several was presenti ng a point of irritation to the clean times blacksmiths ho\'ered in their typical head er, which in this instance was an injury, not a dO\\'ll po:ture before seemingly unresJlon i\'e sea para ite. Some cleaner, for example, Abudel p reh, hut perhaps the sea perch had arlier made dill tl'oschelii, which picks molting skin from the some initialQ'estu re, Whene\'er it could be de Galapa~l.'O , marine iguana (Hobson, 1969b), tl'rmined, the :eaperch initiated the cleaning, will clean dead or injured it sue, but at lea t .'ome dt'taib \\'er ohtained at the :20- to 25-m on this occasion the eaperch sh w d no inter loeat ion a t La Jolin, \\'here two 'eaperch, known est. HOBSON: CLEANI G SYMBIOS IS Material Removed from Other Fish by the perch, 80 or 81 mm long, speared in this habitat, Sharpnose Seaperch were fi ll ed with (showing percent of total vol ume) caprellid amphipods (80% ) , gammarid To determine the food of cleaning seaperch, amphipods (5 % ), isopods (2 % ) , fragments of I examined the gut contents of 16 specimens, algae wi th encrusting bryozoans (10 % ) , and un 74 to 122 mm long, that were spear ed as they identified crustacean parts (3 j!- ). Quast (1968) cleaned blacksmi ths. Food items in their guts, found that specimens from a kelp bed had fed ranked as percentage of each item in the entire mostly on small bottom-dwelling crustaceans, sample, were as follows: caligid copepods, 68ji ; polychaetes and bivalves, as well as kelp frag caprellid amphipods, 16 j( ; gnathiid isopod lar ments, some of which were heavily encrusted vae, 9% ; algae, 1 7c- ; and unidentified items, with bryozoans. Thus the bottom-picking feed 6%. Thus ectoparasitic caligids and O' nathiids ing habits of the white seapel'ch are very similar made up 77 7c- of the material. All 16 specimens to the noncleaning habits of the sharpnose sea contained ectoparasites; in fact, ectoparasites perch. constituted the vast bulk of the mater ial in all On one occasion, I saw a whi te seaperch swim but one individual, which had fed mor e heavil y 1 or 2 m above the surfgrass in company with on caprellids. As with sefi oritas, when an appr e a lone blacksmith, which hovered head-down in ciable amount of free-living material was pres the manner typical of one that desired to be ent, it was usually sharply divided from the ecto cleaned. The white perch picked at the black parasites and more digested to the real' in the smith's body several ti mes, but the bout was digestive tract. All the identifi abl e cali gid cope brief, and the perch soon joined a group of 8 pods among this material were Ca ligus hobsoni, to 10 others of its own kind near the surfgrass '1hich is consistent with what is known of ecto below. This seaper ch, which proved to be 79 mm parasites on bl acksmiths, the species cleaned by long, was speared, and its gut contents included these seaperch, a nd indicates feeding habits si 58 caprell id amphipods, a single gammarid am milar to those of the cleani ng sei'iol'i ta, presented phipod, one small isopod, plant fragments with above. encrusted bryozoans, and some unidentified non parasi tic crustacean remains. No ectoparasites Incidental Cleaning by a Close Relati ve were fo und; its food was similar to that of the Although sharpnose seaj)el'ch were not seen other white sea perch reported above. On an in water less than 20 m deep, the white seaper ch, other occasion I saw a white sea perch cleaning a very similar species, is frequently abundant several blacksmiths over a sandy bottom in 12 m there. The white seaperch was probably the of watel', but this individual was not collected. most numerous of the embiotocids during most Probably the observed cleaning was no more of the observations made at the 3- to 10-m lo than a brier incidental activity for these fish. cation off La Jolla. Underwater the white sea At no other time did I see any indication of clean perch and the sharpnose sea perch a re nearly ing by this species, but perhaps the activity is identical, but can be distinguished by the dusky more frequent under appropriate conditions. bordered caudal fin of the former and the black tipped pelvics and more pointed snout of the CLEAN ING ACTIVITY OF latter. T H E KELP PERCH White seaperch ar e especially abundant in groups of 10 or more close to surfgrass in 3 0 1' Because the kelp perch is not abundant in the 4 m of water off La J oll a. Typically they hovel' La Joll a study area, where larger kelps are head-down; in this attitude they are not soli c spm'se, most observations of cleaning by this iting cleaning but rather are intently r egal'd ing fi sh were made incidentall y during other proj ects the surface of the vegetation, at which they pick in areas heavily fo rested with kelp. However, ::>ccasionall y. Tiny organisms that live on the these other projects generall y were centered on 3urfgrass are prey of these fi sh: fi ve white sea- the sea floor, whereas kelp perch concentrate 513 flsm.RY Bl LLETI. HJL 6? '0 1 aboye in the mid\\'at r and canopy r g- ions. vol LIm ) : g-nalh i id i:oJ od I rva 50 (, ,non Xeyerthel ~~ . ()h~erya ti()n s of cl ,wing- w r suf parasitic isopocls (S',), gammariel amphipods ficiently frequent to recog-nize this. Jl ci • as a (:)',).capr'll!damphipol: (2)', ,anduni(1 n hahitual clean r . though prohahly 1 ss • () than tlflPt\ mat rial (~(J', . .' ith I' of th two ha either thc sennritas or tlll' .1u\enile sharpno ' \\' r not known to hav cl an d con lain d vi ~eaperch. In taking material from th hodi s c! nc of ctopara:it s: on \\a full of eapr lIid of other flshes. the kelp p rch u: s th sam amphipods ( 0',) ancl unid n ifi d mat rial picking techniquc that it emplnys to pick items (10', . \\'h ra. th oth r had nothing in its from an ;Ilgal ~uhstrate . nr that ar adrift in dig- ·ti\ tracl xe pl a f \\' unid n ifi d frag midwater. Its pnintcd ~nout and dcntitinn. which m 'lit: po:t riorly, is ~imilar to that of the sl'llorit:l. a.' descri\)(·d Limhallsrh (19:):» l' port d k Ip p rch clean ahoyc. are \\'ell suited to cleaning. illl! k -Ip l,a .:. opal y :. garibaldi., black. mith , I n~ofar as an aggregation of kelp Ill'rch t nds and \\ allI'Y<' slll'fp ·reh (IfY/lrl'!lrflM)/lOfl a"fl 1/ to remain in one location. thc. e tlsh can p rhap. frllm). he regardcd as maintaining a sta inn at which other flshes are cleaned. nut I ~a\\ no indication that more tlwn one or a few 1l1l'll1h 1': of a g-i\ n :lg-greg:atinn clcan. 01' that other fishes come to \'ariou.· clt'anlng fi:h . r move a wide variety these locations froll1 an~' di. tance for cleaning-. of d I('t l'iou. ma rial from the bodie of the In fact. I ~aw ()nl~' hlacksmith: and other kelp ;lI1lmals th y : J"\·ic. In addition to ectopar 1ll'I'Ch being cleaned hy this l1,.;h. In the nne a. It '. thi: mal rial include. dLea, eel. injured, 'lll~eryatioll of intraspecific cleaning. a sing-Ie "I' necrotic ti ' ~L1e. fungi. anci unwanted food kelp perch s\\all1 among others of it~ aggreg-a particl s (F del'. l~Hj(;; IIob:on. 196 . 1 9b; tion. intentl~' inspecting tlH'ir hodies. t' ually and other:). Howe,' r. th di. cu. ion below con the suhject of this attention mm'ell away. wher - Sid rs cleaninsr only a. the l' moval of ectopar upon the cleaner mm'ed to another tlsh. A few 514 HOBSON : CLEANING SYMBIOSIS interspecific r elations, by affecting not only the seems to be characteristic not so much of a spe relative availability of various prey or ganisms cies as of just certain individuals. At least at and the incidence of various ectoparasites, but a given time, most sei'ioritas do not clean, where also the species composition of the interacting as some seem to be facultative cleaners, and a fishes themselves. few might even be obligate cleaners. Juvenile In California the white seaperch likely is one sharpnose sea perch follow a similar pattern, but of those species t hat cleans only occasionally as with a relatively higher incidence of individuals an incidental adj unct to r egul ar foraging. Sever that clean. Limited data can only suggest that al other California species repor ted by Limbaugh the status of the kelp perch may be similar. (1955 ) and Gotshall (1967) clean, including black per ch, pile perch, and rainbow sea perch, CLEANING INITIATED BY THE SEN-ORIT A but they have not been seen doing so by me. The report of cleaning by the blacksmith (Tur Usually there seem to be fishes present that ner et al. 1969) r emains an anamoly, as this fish need cleaning, as ~ho \\ 'n when a senorita id enti does not fit the pattern of a bottom-picking pred fies itself as a cleaner by initiating action with, ator described above. However, it may be sig say, a blacksmith or a topsmelt, and immediately nificant that many of those substrate-picking is connrged upon by many other fish that crolnl predators which clean most frequently are spe in its way seeking attention, That such fishes cies that also feed on material adrift in mid generally wait for a sei'iorita to begin the clean water , as do the senorita, sharpnose eaperch, ing, rather than attempting to initiate activity and kelp perch. Thus this mode of feeding too, themselns with one of the many sei'ioritas including the taking of plankton, may, in some present, likely refl ects a low probability of suc species, favor adaptations that are suited to cess if they make the first move. If, as it seems. cleaning. Fi hes t hat are adapted to both sub the vast majority of senoritas are not cleaners. strate-picking and plankton-picking may possess or at least not currently predisposed to clean, adaptations especi all y well suited to cl eaning. then random efforts to solicit sen ice would not P robably many species of fishes clean inci seem adaptil'e. dentally on isolated occasions, but relatively few This situation contrasts with that of the Ha are habitual cleaners. And even the habitual waiian wras~e Labmides phthil'opha[Jlls, of cleaner s vary greatly in t hE: degree to which they which all indil'iduals seem to be obligate cleaners are specialized for this habit. Specie of the (Youngbluth, 1968 ) , and which is not nearly Indo-Pacific labrid genus Labroides are highly as abundant on Hal\'aiian reefs as the seilorita specialized cleaners t hat feed almost exclusively is in California. In centering their actil ity on ectoparasitic crustaceans (Randa ll , 1958; around well-defined stations, the disti nctil'e L. Youngbluth, 1968 ) . These fi shes possess many phthil'oJlharfils can be recognized readil~ h~' speci fi c morphological and behavior al specializa others that neeri cleaning. Thus. not surpris tions that are adapted to this way of life (Feder, ingly, clean ing encounters that il1l'ol\'e L. Jlh t h i 1966; Losey, 1971). However, only a small mi I'Opha{Jlls are regularl), initided by fi shes seek nority of cl eaners are so highly specialized; most ing cleaning (Losey, 1971). are but part-time practitioners of the cleaning Vole have seen that under certain circumstallces habit, with much of their food being derived various fishes initiate cleaning encounters from other sources. with seiioritas. Some fishes successfull~' do so That some cl eaners depend on ectoparasites by hovering amid unusually dense concentra for prey, whereas other s can subsist equally well tions of sei'ioritas. but the Ol'eltmes of such I1sh on food from other sources, has led to classifying are not directed at indi\'iduals; rather, they are various species as either obligate or facultative broadcast to the assemblage at large. The suc cleaners (e.g., Youn gbluth, 1968) . The senorita, cess of this tactic presumably folloll's the proba sharpnose seaperch, and kelp perch may well bility that an individual predisposed to clean oc resist being so cl assified because their cleaning curs among such a large number of senoritas. 515 r ISIIERY Bl LLrcTl:-.l VOL (,9, 0 3 Kelpfish regul arly soli cit cleaning from indi all of lhe diffe r nt species are ssentiall y lhe vid ua l senoritas, but lhe siluation is exceptional. same. Because kelpfi sh rise into midwatel' for cleaning, Clean in g by th s 'iiorila may not I speci,s it appears that they do not r eceive satisfactory specific on thos few occa ions wh n the cl eaning service in their regular habitat amid benthic veg activily is initiated hy the fish in need of such etation. In their usual surroundings, wher they service, for example by the kelpnsh, as descri bed arc extremely diflicult lo discem, the cryptic above, Allhough they show. ome difficulty lo kelpfish may be relatiHly inaccessible to clean ing cating a receptiv ~eii()]'ita, kelpfish nev rthel ss sei"\oritas. One can ~ce why a Ii"h thus handi seem far mol' succe. sful at doing so than one capped might be required lo initiale needed would expect if inde d they are required to fi nd cleaning itself. The nu mber of unsuccessful at one that will clean only kelpfi h. Thu, although tempts experienced b~' kelp!lsh he fore a sl'iiorita individual enoritas seem to be specie, -specific was finall y induced to clean them underscores when they them. elves initiate cleaning, they may the existing problem of locating a cleaning indi be considerably less so, and perhaps even non vidual. specific, \\'hen lhe olher fi h make the initial o\'erture. There are no data on thi poi nt, how SPECIE S·S PECIFIC CLEA ING e\'er. The exlent to which these con ider ations ap Because the cleaning senorita initiates most ply to j U\-enile sharpnose eaperch and kelp of its activity, it has the opportunity to elect perch cannot be ascertained because data a re its clients, and the data indicate that a pecies lacking. specific choice is exercised. That individual cleaners tend to limit their . election to member IGNIFICANCE OF PO T RE AS MED of only one species may be related to the fact BY FI HES T HAT OLI IT LEAN ING that they initiate cleaning on the home ground of the fishes they sen'e. when these fishes are When member of an a emblage of fi he like engaged in some of their regular activity. As black ~ miths 01' topsmelt converge on a cleaning each of these clients has distinctive habits, a ei'i site that ha developed in their mid t, probably orita approaching to clean a fish of one species their attention \\'as ini tially alerted by the un faces a somewhat different situation than a sen natural-appearing po ture a umed by the indi orita approaching to clean a fish of another spe \'idual first approached by the cl eaner, sually cies. The distinctions often are subtle, but may thi po tUl'e doe not eem to be a umed pur be significant enough to account fo r a given posefully, but rather results when the fi h, hav sei'iorita's tendency to seek out members of onl y ing cea ed swimming and immobilizing it fins, one species. pas i\'ely drifts out of its regul ar attitude (Hob- Again we can draw a contrast with the clean on, 1965b), The posture thu as umed va r ies, ing behavior of Lab/'oides phthirophaglls, indi especially beb\'een pecie, \\'here perhap dif viduals of which receive members of many dif fering centers of gravity are determining fac ferent species at "'ell-defined cl eaning stations tors. Thus the blacksmith usually hovers head (Randall , 1958; Youngbluth, 196 ). Probably do wn, whereas the topsmelt is more often such nonspecific cleaning is characteristic of tail -down. Sometimes an unnatural-appearing cleaners whose acti\'ity is con fin ed to these estab posture is actively as umed when the fi sh at lished locations. Fishes that \'isit such cleaning tempts to present to the cleaner a certain part stations enter the cl eaner 's own tenito l'Y, and of its body, presumably that part carrying an freq uently join a mi xed-speci es group that hov initation. By virtue of their unusual appear ers in wait for service, In tending these fi shes ance, these postures in cl eaning interactions on its home ground, the cl eaner is receiving serve to dr aw attention to the fi h that is cleaned. them on its own terms, so to speak, so that the It does not eem necessary that any particular situations s urround i n g cleaning bouts with po ture be assumed, only that it look out of the 516 1I0B 01\ : L A:-J ING SnlBIO IS ordinary. Repor ts a re w idespr ead (see Feder, has identitieo it- If by initiating al'ti\'ity with 1966) of cl eaning r ecipients a uming the e un some ti h in the area. Once thL- ha,' occurr d. natural-appearing po tures. one can ee the value of the po.-ture, wh n a _ Attention-gettin O' postures as umed by fi hes sumed by the fir t fi h to be approached. a a being cleaned pr obably occurred incidentally cue in alerting other fi h that need cl aning to during the early development of cleaning sym the presence of a cleaner. In eff ct. then, th biosi , when fi hes hovering to be cleaned quite fish as uming the, oliciting po. tur ac1\' rti . natura ll y stopped m oving and passively drifted the temporary existence of the transient cl aning out of their regul ar attitudes. As the various " tation" to other potential r cipi nl. of cl an cleaning rela tion evolved, apparently this ob ing. Well-oetined cleaning station. like tho: vious cue subsequently assumed a different role of Labl'oides phthil'oll!za{llls 00 not need thi:. ort as a ignal in different situation. Generally of ad\'ertisement, a their location. are \\' II the e posture are suggested to be signals be kno"'n to the fishe that vLit th m. ~or i. it tween the r ecipient of cleaning and the cleaner, necessary that cleaning individual. of L. Jlhtlzi indicatin O' a r eadiness to be cleaned. Quite likely /'Op!l(({/IIS be pointed out. a all m mher. of that this i the prima r y signal-function in activity distindi\'e specie. are cleaners. Despite thi.-. involving such cleaners as Labl'oides phthil'o it is probable that fishes ho\'el'ing to h cl aneo phagus, where a ll members of the species are cit a La/II'oides. tation themselve. create a \'i. ual cleaners an d w here acti\'ity is centered a round cue that tenos to attract other fishe .. cl eaning stations that are well known to other There ll1a~' aL 0 be a maladapti\' aspect to fi he in t he area. I n this situation a fish in the po::;ture::; assumeci b~' tishe. that. olicit clean need of cleaning should be reasonably successful ing. In ho\'ering at an unnatural angle. fin - im in advertising it condition by assllming the mol)ile and erect. a fish may enhance it: chane s characteristic soliciting posture before a fi sh of being cleaned. but it \\'ould also :eem likely recognizable a a cleaner. Losey (1971) hO\\'ed to dra\\' the attention () ]lredator~ ano to handi tha t various fi shes regularly employ this tactic cap iLelf in e\'aoing attack. Perhap: ,uch an to induce L. pht h il'Ol)haglls to clean them. Oh- increaseo yulnerahility accounts at lea. t in part ervations in the Gul f of Califol'l1ia demon. b'ated for the sharp decline in cleaning ('.at occur,' that the cleaning station itself has played a role with reouced visibility. when preoators can ap in establi shing the soli citing posture as a cue. proach clo,er undetected, Increa. eo \'uln ra There I have seen the goatfish Mulloidirhthys bilit~, ma~' also account for the oben'a inll dentaf us hover ing head-down at cleaning ta reported earlier (Hohson. 196:)c) , ,,'here pOll1a tions of the butterfi yfi h H ell iocll1lS nigril'ostris. das~'ios in the Gulf ,f CalIfornia ahruptly hroke when the r e ident cleaner wa itself temporarily a"'ay from cleaner- \\'hen a predator approach eo. ab ent. Lo ey (1971) ob erved imilar behavior among Hawaiian fishes. In such a situation the THE POS IBILITY THAT FI HE BEl ' hovering po ture probably alerts the cleaner to CLEA ED EXPERIEN E TRE fishes that have a rrived f or cleaning. However , in cleaning activity involving the Being prodded and picked 0\' l' by an animal senorita, the oli citing po ture u uall~ ' is assumed of another ~pecie "'ould eem to require a dif only after cleaning has been initiated by the ficult ad.iu~tment for a ti h, It may \\' II he tha clea ner. The problem of recognizing an indi fishes experience. tres, under thL circum tunc . vidual that will cl ean among the va t majority e\'en "'hen the behavior i "'ell e. tahli~hecl. r of enoritas that do not clean. coupled with the tainl~' ohselTation, ha\'e. ho\\'n that hL .'p r ab ence of well-defined cleaning tation, "'ould ience can be uncomfortable, judsring from how reduce the adapti\'ene of the client' oliciting often fi. he. heing cleaned ::;udoenly holt forwarci. po ture a a cue to initiate cleaning. Probably and s\\'im away, apparently ha\'ing h n nil I d the mo t ffective way for a fi h to obtain n eded too \'igorou.ly by the attending clean I' ,001- cl aning in thi situation i to wait until a cleaner time too. a fi h approached by a cleaner cle rly 51; FISHERY BCLLETIN: VOL. 69. NO. J experiences conflicting response, one moment I doubt that cleaners enjoy immunity in tolerating or even soliciting the cl eaner's atten the en~e that predators, recognizing them as tions, and the next moment chasing it away on benefactors, actively a\'oid preying on them. each approach. Such ambivalent behavior was Cleaners may recognize those pr edators which especially evident in rubber lip perch. Losey are not at that time intent on feeding and may re (1971) noted that Labroides phthil'o]Jhaglls in strict their cleaning to such individuals. A preda Hawaii is sometimes attacked by fishes that it tor that assumes a soliciting posture may effec attempts to clean, and suggested that this may t i\'ely ad\'ertise this situation, and no doubt other occur when the cleaning is painful to the host cues exist. uch m chanisms would reduce the fish. chance of cleaners placing themselves in vulner The color changes shown by many fishe being <1ble situations while cleaning. In addition, clean cleaned (Randall, 1958; and others) may in fact cr3 probably are not as vulnerable while cleaning be manifestations of stress. It is \yell known large predators as might be expected simply be that many fishes experience color changes in r e cause cues characteristic of feeding situations sponse to' stress. Earlier (Hobson, 1965a) I dis are not present. In a sociating themselves so cussed the striking color change of the goatfish ir.timately with predators, cleaning fi shes show J!ulloidichthys del/totus ,,'hen it solicits cleaning hehavior that is so unlike that of prey that preda in the Gulf of California, and pointed out that tors probably do not regard them as food. How- this fish sho\ys the same color change in other \'er, e\'en if such factor do reduce the danger situations that are obviously stre, sfu!' Such that might seem inherent in the cleaning act, color changes ha\"e been regarded as signals be I doubt that their cleaning role affords these t,,'een the fishes being cleaned and the cleaners, fishes any ecurity from being eaten in non (e.g., Feder, 1966), functioning in the cl eaning clf'aning situations, interaction much like the soliciting attitude discussed abo\'e. As ,,'ith the attitucles, any role PARASITES AS PREY OF THE such color changes may now ha\'e assumed as a CALIFORN IA CLEANERS signal probably evolyed from an incidental by product of early cleaning. I ha\'e no data on this It is hardly surprising that the fishes which po int i'elating' to the California species, as such are cleaned most frequently in California are color changes are not especially e\'ident in fishes those which are the most abundant and at the that \\"ere observed being cl eaned there. same time carry the most ectoparasites. Thus the blacksmith, topsmelt, halfmoon, and gari ARE CLEANERS IMMUNE FROM baldi are the fi hes cl eaned most frequently, PREDATION? and the survey of ectoparasites showed them to be among the most heavily parasitized. The Reportedly some cleaners are immune from vast majority of ectoparasites on these particular predation because of the service they provide fis hes are mobile forms, mostly caligid copepods the predators (Feder, 1966; and others) . Lim and gnathiid isopod larvae, that occur on the baugh's (1961) belief that the senorita enjoys body surface of their hosts. That these same such immunity is based on observations of this parasites were found to make up the diet of the labrid entering the open mouth of kelp bass to cleaners attending these fishes is consistent with clean and on not finding it among the stomach the observation that only the exteriors of fishes contents of predators during a food -habit study. were seen being cleaned during this study. Howeyer, Quast (1968) found senoritas in the Although the forms infesting the external stomachs of kelp bass, and H. Geoff rey Moser, body surface are the most numerous ectopara U.S. ~ational Marine Fisheries Service (unpub sites on the fi shes available to the California li shed data) , found senoritas in stomachs of the cleaners, many other types were found to infest bocaccio, Sebastes 1JallCispinis, and the starry the oral and branchial cavities. One might ques rockfish, S. constellatus. tion why these other parasites do not seem to be 518 HOBSON : CLEAN ING SYMBIOSIS taken, especially as Limbaugh (1955, 1961) re type predominates; rather, they Occur in a wide ported senoritas entering the mouth of the kelp variety of forms, none widespread among the dif bass and cleaning beneath the gill covers of the ferent species of fishes, and none especially garibaldi. F urthermore, such behavior has been abundant (except on an occasional individual widely reported for some other cleaners, such as fish). Thus a cleaner probably could not subsist species of Labl"oides (Eibl-Eibesfeldt, 1955; on one type alone but would have to master a Randall , 1958; and others) , and some echeneids repertoire of specialized techniques in order to are known to habituall y feed on copepods from exploit enough of these varied forms to make the branchial cavities of sharks (Cressey and it worthwhile. And before access is gained to Lachner, 1970). Nevertheless, any such activity the site of infestation, a much more refined clean by senoritas must be relatiYely rare. In discuss er-host interaction must have evolved than is ing this situation I limit my remarks to the necessary when parasites are simply cleaned senorita, because data are presently insufficient from the external body surface. No such re to determine whether the same may apply to lation would evolve unless the cleaner acquired the sharp nose seaperch and kelp perch. the precise manipalations necessary to pick at Senoritas would not be expected to take par tached parasites off the gills without damaging asites from the oral or branchial cavities as often the delicate gill membranes. Obviously the as species of Labroides or echeneids if for no cleaning relation would not be adaptive if such other reason than they simply are too large rel damage occulTed. In short, to feed habitually ative to most of the fishes they clean. Whereas on parasites from the oral and branchial cavities species of Labl"oides or the echeneids are small would seem to require a higher degree of spe enough to enter the oral and branchial cavities cialization than has been demonstrated by the of most of the fi shes they service, the senorita sei'iorita. It seems unlikely that such special is nearly as large, and sometimes even larger, ization \vould deyelop as long as the more abun than most of its clients. Significantly, Limbaugh dant and readily a\'ailable forms on the body observed senoritas cleaning within the oral and surfaces continue to satisfy the cleaning needs branchial cavities of kelp bass and garibaldis, of the species. Certainly judging from the way both of which are relatively large species. Most blacksmiths, topsmelt, and other fishes vigor of the senorita's cleaning is directed toward ously compete to have their external parasites smaller species, like the blacksmith and the removed, it would seem that there is little imme topsmelt. diate chance of these parasites falling into short The specialized techniques that would be re supply. qui red to prey on the parasites of the oral and branchial cavities would probably pose another CLEANING SYMBIOSIS AND THE problem to the senorita. In its regular habit DISTRIBUTION OF SHORE FISHES of taking parasites from the external surfaces of fi shes, the cleaning senorita concentrates on In his often-cited report on cleaning symbiosis, just a few forms that not only are numerous Limbaugh (1961: 48) stated: on many of the most abundant fishes, but also In my opinion it is the presence of the senorita and are not too dissimilar from free-living prey of kelp perch that brings the deep-water coastal and the species. Sometimes these external forms pelagic fishes inshore to the edge of the kelp beds also occur in the branchial cavity, and some on the California coast. :lIost concentrations of reef similar forms, e.g., bomolochids (Figure 7) , ha fishes ma~' similarly be understood to be cleaning stations. Cleaning stations would therefore account bituall y occur there and in the oral cavity. But for the existence of such weJl-known California sport the majority of parasites characteristic of the fi shing grounds as the rocky points of Santa Catalina branchial and oral cavities are aberrant forms, Island, the area around the sunken ship Valiant off e.g., dichelesthiids, chondracanthids, and lerne the shore of Catalina, the La J olla kelp heds and sub marine canyon and the Coronado Islands. opodids (Figure 7) , and these are unlike 3.11)' thing else encountered by the senorita. Noone Presumably this conclusion was intuiti"e, as no 519 FISHERY B LLETI . VOL. 69, '0 1 data were presented. In his review of cleaning spatial reference point. bviously such fishes symbiosis, Feder (1966: 368), basing his con will c nt r th ms Ives her, b cau e the sur clusion on Limbaugh's work, similarly stated: rounding featureless substrate does not meet "In all probability, many good Ashing grounds their r quirements. I am describing a well are such primarily because they are cleaning known phenom non, one that is the rationale be stations." I believe that this contention is un hind con tructing artiAcial fi hing reefs. Thus founded. Seiloritas are the major cleaners in fo d and a suitable substrate often appear to be California inshore " 'ater , so that if cleaning k y f atures in a habitat that upports large symbiosis does account for most concentrations numbers of A. he. f cour e to cite ju t one or of reef fishes in this region, a Limbaugh ~ ug th other would b an over impliAcation, as re gested, then senoritas would be the cleaner large quirements in both must be atisAed, al ng with ly responsible. Cleaning occ urs wherenl' sen many other perhaps more ubtle need. The var oritas are concentrated but clearly is not a major ious specie as emhled at uch location inter activity of the population, even though it may act in a variety of way ; cleaning symbio is i be so for a relatively few individual. In any one uch interaction, and undoubtedly an impor event, it seems safe to conclude that cleaning tant one, but hardly the prime reason for them is not among the major factors determining the being there. distribution of seil0ritas. And if cleaning does not determine the distribution of senoritas them HA GE HABIT W IT H T IME selves, it seems unlikely that it would determine the distribution of other species. l'ncertaint~· r main regarding changes in Gndoubtedly many factors contribute to cre habib with time. The picture of acti\'ity devel ating situation that draw concentrations of oped in th i: report wa derived di rectly by ob Ashes to certain locations. Where a number of . erYing activity and al 0 indirectly by examining different species have similar requirement, a - both digesti\'e-tract content and the speciAc semblages will deyelop where conditions sati - ectoparasite that infe t the variou A he. But fying these requirements are optimum. The the-e methods only deAne ituation that exi t presence of these Ashes increases the complexity oyer a relati\'ely brief pan of time. Data on of the em'ironment, thus creating situation - that individual activity over longer periods are need support still other species, and so on. Often ed. ertainly habit of individual change with it is apparent that certain features are e pecially time, but how much change and over how much signiAcant as a basis for these concentrations. time? The fact that material in the dige tive Consider, for example, the rocky points that Lim tracts frequently occur in harply delimited baugh included in his li st of "well-known Cali homologou block indicate that the e fi he fornia sport-fi shing grounds." The flora and often feed heavily or even exclu ively on one fauna of these locations are generally rich, a particular type of prey, and then abruptly shift fact probably related to such local features as to omething else. Are habit such a a relative converging current that frequently produce up tendency to clean and to clean members of ju t welling and nutrient-rich waters. Plankton is one species immutable characteristics of indi commonly abundant here, along with plankton viduals, or have the ob ervation described in feeding fishes like the blacksmith. Senoritas this report simply deAned temporary situations and other species frequently are numerous here that the various individual ju t happened to be too, but the main attraction seems to be a gen experiencing at the time they were singled out erally rich food supply rather than available for study? It is pos ible that all senoritas clean cleaning. Similarly it is unrealistic to attribute at one time 01' another, though not all at once, concentrations of fishes around sunken ships to and only a few at a time. It is al 0 po ible that cleaning activity. Where a wreck has settled a senorita, which tend to clean members of j u t on an open expanse it becomes a haven fo r fis hes one specie during a given period of cleaning, that require a nearby structure for cover or a may select members of another species dur ing 520 HOBSON: CLEANI NG SYMB IOS IS a subsequent period of cleaning. Despite these and the same seems to be true of the kelp perch. questions, the conclusions drawn in this report The incidence of cleaners is much higher amonO' are dependent only on an accurate assessment juvenile sharpnose seaperch, but the adults of of the immediate situation, so that their validity this species do not seem to clean r egularly. The is not affected by whether or not the habits of maj or food of all three species is free-livin 0' or individuals under study r emain basically un ganisms which they pick from a sUbstrate'" and changed over time. mid water. 4. There is little overlap between the cleaning A NOTE ON INDIVIDUAL VS. areas of the three species. The sei'iorita is the SPECIES HABITS major cleaner in southern California inshore waters by virtue of its great abundance in a Information on variations in feeding behavior variety of rocky habitats. However, the kelp among individual fi sh under natural conditions perch may be the predominant cleaner in the is difficult to acquire. Typically a given behavior canopy region of the kelp beds, where the spe is described as a species characteristic, and the cies concentrates, and the sharpnose seaperch extent to which this behavior varies among the is the predominant cleaner where it occurs at different members of the species is unknown. depths below about 20 to 30 m and lor water 0 0 Observations of cleaning by the senorita dem under 12 0 1' 13 C, even though the senorita onstrate that different individuals in a popula may be more abundant. tion may react differently to a given situation. 5. The sei'iorita and sharpnose sea perch do not Unquestionably this phenomenon extends beyond establish well-defi ned stations at which they re cleaning behavior to other facets of the animal's ceive other fi shes seeking to be cleaned-a situ activity. If, as is probable, some of the charac ation frequently described for other cleaner teristics of individual fish result from early im fi shes. Rather , as they mO\'e from place to place, printing, then different members of the same individuals of these species approach and clean population could be expected to react differently other fishes in various different locations. in certain situations throughout life. In any 6. Cleaning activity by these species is essen event, it seems unquestionable that the behavior tially limited to removing ectoparasite's from t he of an individual is considerably more limited than external body surfaces of fi shes. They do not that descriptive of its species, or even its own ordinaril y take parasites of the oral and branch population. ial cavities. The dentition of the sei'iorita and kelp perch, which is similar and which includes a number of long, curved canines that project CONCLUSIONS forward at the front of each jaw, seems espe ci ally suited to pick ectoparasites. 1. Three inshore species of fi shes in southern 7. The majol' prey taken by these fishes California are habitual cleaners: the senorita, through cleaning are caligid copepods and gnath the sharpnose seaperch, and the kelp perch. A iid isopod larvae. The species of parasite taken number of other species clean occasionally as most often by the sei'iorita and sharpnose sea an incidental adjunct to t heir r egular feed in g. perch is Caliglls hobsoni. 2. The senorita may clean throughout its post 8. Some species of fi shes are cleaned far more larval life, whereas cleaning by the sharpnose often than others, and many species that co seaperch is an activity largely of juveniles. The occur with these cl eaners are not cleaned at all. life-history period during which kelp perch clean The fishes most frequently cleaned are those has not been defined. which at the same time are most abundant and 3. Cleaning is of secondary significance to most hea\'ily infested with ectopara ites. The these species, a lthough it may be of major sig most numerous ectoparasites on these fishes are nificance to certain individuals. Only a few of cali gid copepods, the most abundant of which is the many sei'i oritas pl'esent at a given time clean, C. hobson i. 521 FISHERY BULL ETIN, VOL. 69, NO.3 9. Cleaning effectively r educes the number of from being eaten during noncleaning situations. ectoparasites that infest the external body sur 18. Cleaners are widespr ead among small faces of fishes that interact with the cleaners. mouthed mar ine fi shes that characteristically 10. At any given time, many individuals of pick tiny organisms from a substrate. This mode the more frequently cleaned species are in need of feeding, especiall y when combined with the of cleaning. Nevertheless, in activity involving capacity to pick tiny prey that are adrift in mid the senorita, infested fishes do not ordinarily water, preadapts fi shes to the cleaning habit. attempt to initiate cleaning but instead wait for 19. There is no basis for the contention that cleaning to be initiated by the cleaner. Because many of the good fis hing grounds in southern the \'ast majority of senoritas are not cleaners, California ar e such because fi shes have congre or at least are not currently predisposed to clean, gated to be cleaned by resident cleaners. random effort to solicit cleaning from se - oritas 20. Feeding behavior varies significantly in the population at large "'ould not be adaptive. among individuals of at least some species. Thus 11. When initiating its cleaning activity, a the habits of an individual can be more limited gi\'en individual senorita tends to approach and than those descriptive of its species or even its clean members of only a single species of fish. own population. 12. Because the vast majority of senoritas are not currently predisposed to clean and because there are no well-defined cleaning stations, the unnatural-appearing posture assumed by a fish ACKNOWLEDGMENTS approached by a cleaner is an important cue in adverti ing the location of available cleaning to Lloyd D. Richards assisted me in collecting other fish in need of this senice. the data for this study. I thank Roger F. Cressey 13. Fishe being cleaned probably experience and Thomas Bowman, U. S. National Museum, some degree of stress. The color changes ex who identified the parasitic copepods and iso hibited by some fishes when being cleaned are pods, respectively, ar ound which much of this essentially manifestations of this stress; sec study is centered. I thank, too, Carl L. Hubbs ondarily, they may ha\'e assumed a signal-func and Richar d H. Rosenblatt, Scripps Institution tion in certain cleaning interactions. of Oceanography, and J ohn R. Hunter, U.S. Na 14. \Yhi le intimately associated with the fishes tional Marine Fisheries Service, for helpful com they clean, senoritas frequently become infested ments on a draft of the manuscr ipt. themsel\'es by the same parasites they are at tempting to remove from these other fishes. 15. Cleaning acti\'ity is sharply curtailed when visibility is reduced by turbid water or when there is strong water movement, such as a heavy REFERENCES surge. CLARKE, T. A. 16. Cleaning activity among these fishes is a 1970. Terr itorial behavior and population dynam diurnal phenomenon. There is no evidence t hat ics of a pomacentrid fi sh, the garibaldi, H ypsypops it continues after dark. rHbiclluda. Ecol. Monogr. 40 : 189-212. 17. Any so-called "immunity" from predation CLARKE, T. A., A. O. FLECHSIG, AND R. W. G RIGG. that a cleaner may enjoy probably relates (1) to 1967. Ecological studies during P r oject Sealab II. Science (Wash., D.C.) 157 : 1381-1389. an ability to recognize predators that are not in CRESSEY, R. F . tent on feeding and to limit cleaning to such indi 1969a. Caliglls hobsoni, a new species of pa r a sitic viduals, and (2) to the fact that behavior exhib copepod f rom Ca lifornia. J . P a r asitol. 55 : 431- ited by a cleaner senicing a predator is so unlike 434. that of prey that the predator does not regard the 1969b. Five new pa r asitic copepods from California inshore fi sh. Proc. BioI. Soc. Wash. 82: 409-427. cleaner as food. However, their r ole as cleaners 1970. H atschekia pacifica new species (Copepoda: probably does not afford these fis h any security caligoida ) a parasite of the sand bass, P aralebrax 522 HOBSON : CLEANING SYMBIOSIS n ebuli/ er (Giard). Proc. BioI. Soc. Wash. 82: L ONGLEY, W. H ., AND S. F . HILDEBRAND. 843-846. 194 1. Systema tic catalogue of the fishes of Tor CRESSEY, R. F ., AND E. A. LACH NER. tugas, Florida. Pa p. Tortugas Lab. 34. Carnegie 1970. The parasitic copepod diet and lif e history Inst. Wash. Publ. 535, 331 p. of diskfishes (Echeneidae). Copeia 1965: 310-318. L OSEY, G. S., JR. EIBL-EIBESFELDT, I. 1971. Communication between fishes in cleaning 1955. Dber Symbiosen P arasitismus und andere symbiosis. I n T. C. Cheng (editor ), Aspects of besondere zwischena rtliche Beziehungen tropis the biology of symbiosis, p. 45-76. Univ. Park cher Meersfische. Z. Tierpsychol. 12 : 203-219. P ress, Baltimore. FEDER. H. M . MILLER, D. J ., D. G OTS H ALL, AND R. NlTsos. 1966. Cleaning symbiosis in the marine envi ron 1965. A fi eld guide to some common ocean spor t ment. In S. M. Henry (editor) , Symbiosis, Vol. 1, fi shes of Calif ornia. 2d revision. California De p. 327-380. Academic P ress, New York. partment of F ish and Game, 87 p. GOTSHALL, D. W. QUAST, J . C. 1967. Cleaning symbiosis in Monter ey Bay, Cali 1968. Observations on the food of the kelp-bed f ornia. Ca lif. F ish Game 53: 125-1 26. fi shes. In W. J. North and C. L. Hubbs (editors), Utilization of kelp-bed resources in southern Cal H OBSON. E . S. ifornia, p. 109-142. Calif. Dep. Fish Game, Fish 1965a. F orests beneath the sea. Anima ls 7: 506- Bull. 139. 511. R ANDALL, J. E. 1965b. A visit with el bar bero. Underwater Nat. 1958. A review of the labrid fish genus Labroides, 3 (3) : 5-10. with descriptions of two new species and notes on 1965c. Diurna l-nocturnal activity of some inshore ecology. Pac. Sci. 12: 327-347. fishes in the Gulf of California. Copeia 1965: 1962. Fish service stations. Sea Frontiers 8: 291-302. 40-47. 1968. Predatory behavior of some shore fi shes in ROEDEL, P. M. the Gulf of California. U. S. Fish Wildl. Serv., 1953. Common ocean fishes of the California coast. Res. Rep. 73, 92 p. Calif. Dep. Fish Game, Fish Bull. 91, 184 p. 1969a. Comments on cer ta in recent genera li zations SHII NO, S. M. regarding cleaning symbiosis in fi shes. P ac. Sci. 1956. Copepods parasitic on Japanese fishes. 12. 23: 35-39. Family Lernaeopodidae. Rep. Fac. Fish. Prefect. 1969b. Rema rks on aquatic habits of the Galapagos Univ. Mie 2: 269-311. marine iguana, including submergence t imes, TURNER, C. R., E. E. EBERT, AND R. R. GIVEN. cleaning symbiosis, and the shark threat. Copeia 1969. Man-made reef ecology. Calif. Dep. Fish 1969: 401-402. Game, Fish Bull. 146, 221 p. H UBBS, C . L., AND L. C. H UBBS. WILSON , C. B. 1954. Da ta on the life history, varia tion, ecology, 1917. North American parasitic copepods belonging and rela tionships of the kelp perch , Bmchyistius to the family Lernaeidae with a revision of the /renatus, and embiotocid fi sh of the Californias. entire family. P roc. U.S. Natl. Mus. 53: 1-150. Calif. Fish Game 40: 183-1 98. 1935. Parasitic copepods from the Pacific coast. LIMBAUGH, C. Am. MidI. Nat. 16: 776-797. 1955. Fish life in the kelp beds and the effects of YOUNGBLUTH, iVI. J . kelp harvesting. Univ. Calif. I nst. Mar . Resour. 1968. Asp ect~ of the ecology and ethology of the IMR Ref. 55-9, 158 p. cleaning fish, Labroides phthirophagus Randall. 1961. Cleaning symbiosis. Sci. Am. 205 : 42-49. Z. Tierpsychol. 25: 915-932. 523